Study and Lexicogrammatical Proposal of Physical Perception Verbs in English

FACULTAD DE FILOLOGÍA DEPARTAMENTO DE FILOLOGÍA INGLESA (LENGUA INGLESA) DOCTORADO EN ESTUDIOS FILOLÓGICOS STUDY AND LEXICOGRAMMATICAL PROPOSAL OF PHYSICAL PERCEPTION VERBS IN ENGLISH Tesis presentada para la obtención del título de Doctora en Filología por: Mª DOLORES LAGIER REINA Tutor/director: DR. D. JOAQUÍN COMESAÑA RINCÓN Co-director: DR. D. FCO. JAVIER TAMAYO MORILLO Sevilla, febrero de 2025 i Abstract The aim of this doctoral dissertation is twofold: a) to study one of the most relevant verb classes, the physical perception verb class, and b) to put forward a proposal for its lexicographical or, following Comesaña Rincón et al (2013), lexico-grammatical description. Despite the fact that this verb class has received substantial attention, there are several aspects that are still overlooked or undefined and, thus, not properly reflected in dictionaries, e.g., its syntactico-semantic potential. Hence, the current study is divided into two main sections: first, a theoretical characterization of physical perception verbs in English, and second, an applied characterization of such verbs. In the first part, a semantic classification of the members of this verb class will be established, serving as the starting point for their latter examination. This way, their syntactic and semantic complementation patterns, their use in the progressive structure, as well as their nature as carriers of evidential meaning or their behavior as DMs, among other aspects, will be explored. Furthermore, how these verbs are usually treated in some of the major dictionaries will also be described. The second part, on the other hand, will consist in the proposal of a lexicographical description of these verbs in a bilingual (English-Spanish) dictionary-like form. To do so, the different meanings that these verbal forms can convey will be collected and analyzed according to all the information discussed in the first part of this dissertation. Altogether, this proposal will offer a deeper and more comprehensive characterization of this verb class, covering aspects typically absent in dictionaries. Keywords: DMs, evidentiality, lexicography, metaphorization, physical perception verbs, semantic characterization, stative verbs, syntactic complementation. ii iii To my family, thank you for listening. iv v Acknowledgements The process of writing a doctoral dissertation is highly demanding and sometimes can be perceived as lonely as well. I have been fortunate enough to have the help, support, and encouragement of many people around me. It is thanks to them that the completion of this doctoral dissertation has been possible and that I have been able to see through this long yet not lonely journey. First of all, I am highly indebted to my two supervisors, Joaquín Comesaña Rincón and Fco. Javier Tamayo Morillo, for their unwavering generosity, patience, availability, and insightful comments and advice. They have always been an inspiration and an example to me. I will not be able to thank them enough for their continuous academic and, above all, personal support throughout these years. My gratitude also goes to Elaine Uí Dhonnchadha for being my supervisor at Trinity College Dublin and giving me the opportunity to carry out a research stay there, as well as to Valentina Colasanti and Daniela Modrescu. Not only did they three helped me academically, providing me with resources, comments, and advice, but they also made me feel at home. It was a privilege to share that time with you. I would also like to thank Ángela del Estal Poveda, who was always there for me since the beginning, helping me find the right track in this new academic world with enormous patience, generosity, and care. Likewise, I want to thank the members of the English Language Department, specially Ángel L. Jiménez Fernández for all the trust and opportunities, Julia Fernández Cuesta for all her work and effort as part of the Academic Committee in the PhD program, Cristina Molinos Rodríguez de Trujillo for her support and feedback, and Isaiah O’Bryon for his help with some of the English examples and lexical entries. Though no longer part of this Department, I would also like to mention Francisco Garrudo Carabias, whose work served as the stepping-stones of this research. I am also grateful to Marta León Prol, my friend, fellow PhD candidate, and colleague, for her kindness, encouragement, comments, and, of course, for listening to my “podcasts.” To my old friends Myriam and Raquel, for those board game evenings that helped me relax and for being always there for me without constantly asking me “when will you be done?” And to Laura, for her unconditional friendship despite not being able to see each other as often. Most importantly, this thesis would not have been possible without the encouragement and trust of my family. I would like to thank them for putting up with me in those times when I felt lost and frustrated and, most importantly, for believing in me during this long process vi even when I did not have the confidence to believe in myself. They have always encouraged me to keep on going. I could not have done this without their unfailing love and support. To them I dedicate this dissertation. Finally, this doctoral dissertation has also been possible thanks to the financial support of the V Plan Propio de Investigación of the University of Seville (Spain) through a predoctoral research fellowship (Contratos Predoctorales o de Personal Investigador en Formación (PIF) para el Desarrollo del Programa Propio I+D+i de la US, VPPI-2016-II.2) and a grant for research stays (PP2019-ES041). The research here has also been partially funded by the I+D project PID2022-137233NB-I00 granted by the Spanish Ministry of Science and Innovation (MICINN). These grants are hereby gratefully acknowledged. vii CONTENTS Abstract ....................................................................................................................................... i Acknowledgements ..................................................................................................................... v List of Figures ........................................................................................................................... ix List of Tables ............................................................................................................................. ix CHAPTER 0. Introduction ...................................................................................................... 1 CHAPTER 1. Physical Perception Verbs .............................................................................. 7 1.1. The predicative unit par excellence: The verb ................................................................. 7 1.2. Perception and language .................................................................................................. 9 1.3. The physical perception verb class in English ............................................................... 14 1.4. Grammar and dictionaries ............................................................................................. 16 1.4.1. Grammatical information in dictionaries ................................................................ 21 1.4.2. Syntactic and semantic aspects of complementation in dictionaries ...................... 26 CHAPTER 2. Semantics ........................................................................................................ 35 2.1. The pivotal element: The meaning ................................................................................ 35 2.2. Semantic classification of physical perception verbs .................................................... 40 2.3. Semantic characterization .............................................................................................. 50 2.3.1. Two arguments ....................................................................................................... 53 2.3.2. One argument .......................................................................................................... 57 2.3.3. More than two arguments ....................................................................................... 58 2.4. The semantic domain of evidentiality ........................................................................... 58 2.4.1. Evidentiality and epistemicity ................................................................................ 63 2.4.2. The evidential meaning in physical perception verbs ............................................. 67 CHAPTER 3. Syntax .............................................................................................................. 73 3.1. The progressive aspect ................................................................................................... 73 3.1.1. Stative vs. Dynamic verbs ....................................................................................... 73 3.1.2. Physical perception verbs and the progressive: Their interaction .......................... 76 3.2. Syntactic complementation ............................................................................................ 82 3.2.1. Intransitive .............................................................................................................. 84 3.2.2. Transitive ................................................................................................................ 86 3.2.2.1. Monotransitive ................................................................................................. 86 3.2.2.2. Ditransitive ....................................................................................................... 87 3.2.2.3. Complex Transitive .......................................................................................... 88 4 Secondly, it provides a semantic classification of physical perception verbs along the lines of Garrudo Carabias (1999: 185; 2015), whereby not only is each of the five senses taken into consideration, but also the specific semantic role of the Subject-referent. Thirdly, the different semantic frames which can be found in this verb class are introduced. And lastly, the notion of evidentiality is also addressed, for these verbs can be considered as the most representative evidential markers in English. Chapter 3. presents an account of mainly syntactic aspects. “Syntax is concerned with the way words combine to form phrases, clauses, and sentences” (Huddleston & Pullum 2012: 4). The syntactic features related to the Verb Phrase in general —and to the physical perception verb class in particular— have been largely treated in the linguistic literature. Hence, this chapter only discusses three aspects of syntax. It examines the long-established classification of physical perception verbs as a stative verb class, pointing out their heterogeneity in this respect and the crucial role played by the specific context, syntactic and pragmatic, in determining their semantic reading and, in turn, their stative or non-stative (or dynamic) categorization. In a similar fashion as it was done with the semantic characterization, the various syntactic models of complementation which these verbs can show are also discussed, showing their wide diversity. Finally, this chapter also tackles the frequent use of certain physical perception verbs as discourse markers, paying attention to the distinct communicative functions they serve and their potential with respect to the information structure of the utterance. This topic that has yet to receive enough scholarly attention. Chapter 4. begins by briefly examining two approaches regarding lexical ambiguity, namely the traditional —and not so clear-cut— distinction between polysemy and homonymy. Then, it focuses upon all the different strategies that take part in the semantic extension of verbal units, particularly, of those under study. This way, this chapter analyzes the syntactico-semantic processes that participate in this verb class as well as metaphorization, which is considered the most productive mechanism as far as semantic extension is concerned. Chapter 5. exposes the methodology employed for the creation of our lexicographical proposal of these verbs. It includes the criteria followed regarding what constitutes a separate lexical entry, the information that should appear in it, and its internal layout, i.e., its microstructure. In addition, other aspects related to this proposal, such as the notational system and the examples included, are also described. Chapter 6. constitutes a major part of this doctoral dissertation: the proposal of a lexicogrammatical description of the most frequent members of the physical perception verb class. Here, the verbs studied throughout this investigation are presented in a bilingual 5 dictionary-like form (English-Spanish) which encapsulates those factors previously analyzed, based on previous work by Garrudo Carabias (1991, 1996) and Comesaña Rincón et al (2013). Finally, Chapter 7. outlines the main conclusions to be drawn from this study in relation to all the diverse topics discussed. Furthermore, it also offers some suggestions for further research, some of which we intend to carry out in the near future, as well as some possible applications that this study, and, in particular, the proposal herein, has in other subfields of Linguistics. 6 7 CHAPTER 1. Physical Perception Verbs 1.1. The predicative unit par excellence: The verb It has been long established that among all the predicative units, namely nouns, verbs, adjectives, adverbs, and prepositions, the verb sticks out in a privileged position. It has received the greatest attention in the linguistic literature and has been widely studied over time by authors galore, not just in English, the so-called ‘global language’ or ‘lingua franca,’ but in many other different languages as well. Likewise, the verb has also been the focus of numerous crosslinguistic studies regarding languages pertaining to the same or different families, refer to Comrie (1976), Ibarretxe-Antuñano (1999, 2002), Usoniene (2001), Rojo and Valenzuela (2004), Levin (2008), Viberg (2008), Whitt (2010), or Neagu (2013), among others. Contrarily, the rest of predicative units has been treated as Cinderellas and, comparatively speaking, their research being much too limited and sparse. 1 The lack of balance around the quantity of research on the different predicative units reflects a pattern that is still followed nowadays and, although this highly verb-focused tendency may not seem ideal, there are sound reasons behind it. On the one hand, the verb is the most central element in a sentence and controls intrasentential syntax, i.e., the sentence is structured around the syntactico-semantic behavior of the verb. Hence, it is the one determining both the number and nature of the complements/arguments that accompany it (also called valency in Valency Grammar, see Herbst 1987 and Herbst et al 2004), 2 as well as their position within the sentence. For this very reason, and as Comesaña Rincón (2000: 206) points out, “grammatically, sentence correctness depends on the different elements required or rejected by […] the verb.” That is to say, despite a sentence showing the correct word order, appropriate tense, and agreement among its elements, that sentence would still be considered grammatically incorrect or ill-formed if the syntactico-semantic qualities of the elements appearing in it do not match with the ones licensed by the verb and the other way round. On the other hand, in all types of contexts of Language Acquisition, that is, First, Second, and Foreign Language Acquisition, verbs and their syntactico-semantic features are more difficult to learn than nouns, for instance. As Saeed and Fareh (2011: 4) state: 1 Let us just mention, for example, Rusiecki (1985), Comesaña Rincón (1987, 1989), Ferris (1993), Bhat (1994), and Tucker (1998) regarding the adjective, and Herbst (1988) and Tamayo Morillo (1998) with respect to the noun. 2 A terminological note. In the present investigation the notion of argument will be used to refer to those elements that are necessary for the syntactico-semantic completion of the predicative unit, in this case, the verb, without it being restricted to either the syntactic or the semantic level of analysis. 8 This difficulty may be attributed to the fact that verbs are associated with complex semantic and syntactic features pertaining to tense, time, aspect and phase. Thus, verbs are cognitively more demanding than nouns in learning (Gentner 1981, 1982, 2006; Gropen et al 1991; Kersten – Smith 2002; Mandler 1996). There is great variability in the morphological system of English verbal units. 3 Furthermore, the number of irregular forms is also larger in verbs than in any other predicative unit. This means that learners will be exposed to a higher degree of variation regarding the appearance or surface morphological form of verbal units. Semantically, verbs also encapsulate more abstract information and diffused relations. According to the natural partitions and the relational relativity hypotheses (Gentner 1981, 1982, 2006; Gentner & Boroditsky 2001), the mapping from language to the world, i.e., from word to referent, is less transparent in the case of verbs and, in addition, varies substantially cross-linguistically (refer to Talmy (1975) or Slobin (1996)), increasing its learning difficulty. Apart from that, the participation of several syntactico-semantic processes and other factors that influence or modify the meaning of the verb in different ways —as well as its behavior and syntactico-semantic configuration in some instances— also squarely contributes to an increment of the number of senses and, thus, of complexity (for further discussion, refer to Chapter 4.). Taking all this information into account, verbs can be certainly described as extremely complex predicative units at all different levels of linguistic description. Given their relevance and complexity, verbs have been fully analyzed in various fields and, naturally, divided into classes. These are mostly, though not exclusively, based on semantic grounds, more specifically, on their meaning since “further examination of the nature of lexical knowledge confirms that various aspects of the syntactic behavior of verbs are tied to their meaning” (Levin 1993: 5). From animals to plants, geometrical figures or musical instruments, categorization is essential for cognition, that is, to bring about some order and structure that enables us to understand more effectively and thoroughly the world that surrounds us. The aim of establishing these verb classes is, then, to contribute to structure the broad verb lexicon and organize its many elements by highlighting linguistic relations among them, 4 thus allowing a better description of their meaning and behavior. Besides, as Fillmore (2003) signals, verb classes could also be used to identify grammatically relevant components of verb meaning. 3 Refer to Quirk et al (1985: 96-120; 127-137) for a comprehensive classification of verbal units according to the paradigmatic formal variants and the different formal subtypes. 4 We are not unaware that this label can have different interpretations depending on the author as well as on the linguistic trend adopted (for instance, Generative Grammar). In the current study, the term ‘lexicon’ will be broadly understood as vocabulary. 9 There are plenty of verb classifications and subclassifications. They all depend on which perspectives are followed and/or which criteria are considered and given prominence to when analyzing the verb repertoire. One of the most valuable classifications is the one made by Levin (1993), who distinguishes 48 semantic verb classes plus their subsequent subclasses according to their participation in 79 syntactic alternations involving mainly NP and PP complements. In the present investigation, the focus of attention will be on one particular verb class: the physical perception verb class. Interestingly, this specific class is not included as such in Levin’s classification. Instead, she groups these verbs into several different classes: verbs of contact (1993: 155-156), verbs of perception (1993: 185-188), verbs of smell emission (1993: 236237), or verbs of sound existence (1993: 252-253), to name a few. 5 This fact serves to highlight both the heterogeneity within the members of this verb class as well as the heterogeneity that exists among the multiple verb classifications. 1.2. Perception and language Perception has been described as the process by which a stimulus is detected through one or more senses (sight, hearing, smell, taste, and touch), decoded by the corresponding organ(s) (eyes, ears, nose, tongue, and skin, respectively), and interpreted by the brain, which transforms it into meaningful information and elaborates a response. 6 Perception represents a biological process with an outstanding relevance in our lives, that is why it bears an influence on many different areas such as Psychology, Philosophy, and, of course, Linguistics. Perception has always been used as a method for survival by human beings and by any living being in general. Following Darwin’s (1859) evolutionary theory of natural selection, species have adapted themselves and, consequently, their senses to their own environment in order to survive. As an example, animals such as birds have developed an extraordinary sense of vision, they have a wider range of view than many other animals, and excel at perceiving small motions and details, which enables them to find food as well as to detect and evade stealthy predators more easily. 5 For instance, touch is included within verbs of contact; see, hear, sniff or taste under verbs of perception; smell and stink are classified as members of the verbs of smell emission; and sound as a verb of sound existence. 6 As noted by Winter (2019: 12-15), this five-senses model is not without issues. It has been described as culture-specific by anthropological research (Howes 1991) and scientists recognize other sensory systems within our physiology for sensing, for instance, pain (nociception), balance and spatial orientation, and our body’s position and self-movement (proprioception). The high degree of interactions between senses also poses another question concerning the actual boundary between them as well as their categorization. For this research, however, we will follow this five-senses model as it has been proved to be appropriate for studying the English language and, especially, its physical perception lexicon (Norcliffe & Majid 2024). 10 With respect to human beings, we have constantly relied on our senses to perceive what is around us, to examine our environment, and to identify not only any potential danger we may encounter, but also what could be beneficial to (or useful for) our lives. We use our senses to orient ourselves when we are lost or somewhere unknown, to find out if something is edible, or even to communicate —verbally or non-verbally— with other people. Even philosophers as renowned as Aristotle emphasize their primacy as the starting point of any form of knowledge. As the peripatetic axiom expresses, nihil est in intellectu quod prius non fuerit in sensu [nothing is in the intellect that was not first in the senses]. 7 In this way, the five senses are used as “our communication channels to reality” (Sekuler & Blake 2006: 3) and, what is more, perceptual information plays a decisive role during young children’s mapping process from world to language and vice versa. As each sense is linked with a separate organ, they respond to different stimuli and provide different kinds of information, making it possible for us to use one sense in isolation or various together depending on our situation and needs. That being said, despite the fact that some senses can be used and blocked at will in certain circumstances (we can close our eyes or cover our ears), most of the times we have no power of decision over what we perceive. The main reason is that we are constantly being exposed to different types of stimuli from the moment we wake up and even while we are sleeping, which makes it almost —if not completely— impossible to always control which stimuli we want to receive or reject. Besides, some of these stimuli come rather abruptly and, even if we wanted to, we would have no time to react properly. Let us think, for instance, of the sound of a firecracker or of a car horn. 8 A central element in the process of perception is, indeed, the stimulus (also known as percept). The stimuli we may perceive in our daily life are quite varied, both qualitatively and quantitatively speaking (Fernández Jaén 2012: 157). Firstly, because of the large diversity 7 This statement is not universally acknowledged by any means. Empiricists, such as Hume or Locke, do assert that our source of knowledge of the external world is, indeed, experience, gained through our senses. Nonetheless, rationalists, e.g., Descartes or Kant, posit that knowledge is gained through intuition and reason, and hold the opinion that we are born with an a priori knowledge as part of our rational nature, which is not provided by experience, although it is experience that triggers our awareness of it. This is known as the Innate Knowledge Thesis (Markie & Folescu 2021). 8 Three senses are particularly curious in this regard, namely sight, smell, and touch. We can constantly perceive our own nose through our sense of sight, but when that information is transmitted to the brain, this organ decides to ignore it as it is always there, does not provide us with much information, and blocks part of our field of vision. Something similar happens with our senses of smell and touch. After being exposed to the same smell or to a stimulus warmer/colder than physiological zero for a certain prolonged period of time, we get used to them and our brain decides to stop perceiving those stimuli. It seems as if we partially or momentarily lost those precise senses. This is known as sensory adaptation in psychobiology, i.e, “a temporary reduction in the responsiveness of a sensory receptor other than a pain receptor as a result of repeated or continuous stimulation” (Colman 2015: 11). 11 among the five senses the source of the stimuli we can experience or perceive through them can be either external or internal. That is to say, the stimulus can come from the external physical world as, for instance, the heat of a fire, the sound of an alarm, or the smell of a flower, or from our own body, like pain, tiredness, or hunger. Secondly, the frequency with which we perceive these stimuli also varies. Certain stimuli can be normally found in our daily life while others are deemed more exceptional or rare. It is more common overall to hear the sound of the birds chirping than to feel seasick, or to see a cloud than to experience pain because of a broken bone. Finally, there is yet another aspect which is built on more subjective values: our opinion about them. Whether we find a specific stimulus pleasant or unpleasant is mostly a matter of subjectivity or taste. Scents are a clear example of this, some people may like more floral scents, whereas others may dislike those and prefer fresh ones instead. Likewise, an individual may find traffic noise stressful and annoying, whereas another may think of it as something relaxing, similar to ambient sounds. It should be noted, however, that although this aspect is mainly subjective, there are also some objective examples that everyone —or at least, most people— seems to agree on. As an illustration, we can mention the case of pain, which is generally thought of as something unpleasant and that should be avoided. It is worth pointing out that in spite of the universal status of the process of perception in our lives, the level of importance of each specific sense modality is not universal by any means. This statement is a topic of dispute among many linguists, creating a dichotomy between what is known as universalism and relativism. Some scholars hold a universalist view, that is, they assert that there is a common rank ordering of the five sense modalities that remains unaltered across languages or cultures. In support of this viewpoint, Viberg (1984) carried out a seminal typological study which included a questionnaire translated into 53 languages and, based on its results, he established a hierarchy of the five sense modalities that could be applied universally. This hierarchy places sight as the predominant sense, followed by hearing and touch, and relegates the senses of smell and taste to a shared final position: sight > hearing > touch > smell taste Figure 1. Viberg’s (1984: 136) hierarchy of sense modalities. Despite the fact that, as Viberg (1984: 124) himself acknowledges, his language sample is not balanced, as there are “highly underrepresented” and overrepresented languages, his hierarchy has received the approval and consensus of several authors (refer to Sweetser (1990), 12 for instance). On the opposite side, there is also a group of scholars who reject such universalist claims and state that the rank of the sense modalities will differ from language to language and from culture to culture. This non-universalist or relativist view is founded upon the idea that “sensory perception is a cultural, as well as a physical act” (Classen 1997: 401) and, of course, cultures may differ, as in many other aspects, in their own conception of the five senses, varying the value and salience they attribute to each particular sense modality. Therefore, according to relativist hypotheses, cultural variation will lead to the creation of diverse hierarchies of the five senses, as opposed to the universalist premise that holds that there exists one single and generalized hierarchy for all languages and cultures. These two stances, the universalist and the relativist, should not be understood as being mutually exclusive. That is to say, the fact that one of them is considered right in a certain aspect does not necessarily imply that the other is completely wrong in the rest. Actually, they each reflect a part of the reality of physical or sensory perception. On the one hand, a fact that coincides with the universalist theories is that sight seems to be the dominant sense in every culture as recorded in numerous studies, for instance, those by Viberg (1984), Evans and Wilkins (2000), San Roque et al (2015), and Norcliffe and Majid (2024). But, on the other hand, it has also been demonstrated that the rest of the senses do not always occupy the same hierarchical position across cultures, revealing differences of salience as relativists explain (San Roque et al 2015). In Evans and Wilkins’ (2000: 546) words, “the same domain can have its universal and relativistic sides; a foot in nature and a foot in culture.” Physical perception encompasses, in this way, both universalist and culture-specific or relativist sides. Focusing on the universal salience of sight, various reasons could be signaled as the motivation for it. Sweetser (1990: 39) pointed out that sight is “our primary source of objective data about the world.” This assumption is, in turn, grounded on previous studies on two separate areas: child language and evidentiality. Considering the former, during early childhood it is our vision that has the most marked features involved in concept formation (Clark 1976), for not all objects produce sound and children cannot (and, actually, should not) smell, taste, or touch everything around them. With respect to the latter, “crosslinguistic studies of evidentials show that direct visual evidence is considered the strongest and most reliable source of data” (Sweetser 1990: 39), hence the saying seeing is believing (this will be further discussed in 2.4.). Beyond —or besides— Linguistics, there is yet another field that also favors this salience, namely Anthropology. Different arguments can be used to support this. Humans rely more on vision than on any other sense. We use this sense the most when interacting with the world, because it is highly developed. And our brain is more engaged in processing information 13 acquired through vision than through the rest of the senses, as that task involves a large region of the human neocortex (Pike et al 2012: 67). As San Roque et al (2015: 4) expose: The primacy of vision over the other senses appears to be well supported by our underlying biology. Humans, like other primates, display considerable visual specialization including high visual acuity, stereoscopic vision, trichromacy, and large visual cortices (Barton 2006). Some estimates suggest that up to 50% of the cortex may be involved in visual function (Palmer 1999). Experimental studies also support the dominance of sight over auditory perception (e.g., Colavita 1974; Spence 2009). Therefore, the apparent linguistic dominance of vision may be rooted in our pan-human evolutionary history. The stimulus can also be named as a contributor to the “primacy” of sight. Due to the own specific nature of each of the senses, there are certain pre-established conditions that must be satisfied by the stimulus in order to be perceived. For an object or entity to be perceived through hearing, it must have the indispensable capacity to produce sound, to perceive a stimulus through our sense of taste, we must put it inside our mouth, and if we want to smell or touch something, we have to be close enough to it. In comparison to these senses, sight is “selectionally much less restrictive” (San Roque et al 2015: 20), as it does not necessitate any of those requirements from its potential stimulus, the only condition is that the object or entity must be visible, 9 which most daily objects or entities are. In sum, the types of stimulus that can be perceived through vision represent a broader category than those of the sense modalities lower in Viberg’s hierarchy. There is an aspect unique to humans that sets us apart from the rest of organisms and that is closely connected to the process of perception: language. One of the main functions of language, as defended by Cognitive Linguistics (Cuenca & Hilferty 1999: 19), is to describe the world around us and express how we, by means of our senses, perceive it. Our reality and the way we see it differs from culture to culture as well as from individual to individual, since it is based on our own subjective experiences and knowledge of the world, not to mention that the world itself is steadily evolving. In the same way, language also varies depending on our social and cultural environment and is continuously evolving and adapting itself to appropriately depict new realities. Language acts, then, as a mirror and as an embodiment —as known in Cognitive Linguistics— of our world, and, as a result, the general pre-eminence that perception has in life is clearly reflected in the lexicon, occupying a salient position within it. 9 Let alone metaphorical sight, which will be dealt with later on. 20 2001: 97). Echoing Garrudo Carabias’ (2003: vi) words with respect to the nature of bilingual dictionaries: Un diccionario bilingüe mal diseñado, concebido como simple repertorio de lexemas sobre los que se da una información meramente léxico-semántica, es una hoja de doble filo, y sin mango. Un buen diccionario bilingüe, por el contrario, es una navaja suiza con múltiples aplicaciones: debe ayudar a descodificar o interpretar, y a codificar o producir en cada una de las dos lenguas implicadas, informando de manera precisa sobre circunstancias colaterales de tipo fonético, sintáctico, pragmático, de registro y uso; y tratando de formalizar o exteriorizar el conocimiento ideal e interiorizado, existente en el lexicón mental de un hablante nativo. 21 This scenario described by Garrudo Carabias, which can be certainly applied to monolingual dictionaries as well, implies two different but connected ideas. Firstly, as much as the dictionary is an extremely useful tool, it can also become quite dangerous if it is not welldesigned, as it would leave the door open for users to commit errors despite its appropriate use. And secondly, the acknowledgement that, as Abad Nebot (1997: 90) quoting Lázaro Carreter (1980: 74) also expresses, dictionaries have a double purpose for learners: their traditional function as “instrumentos descifradores en lo fundamental,” 22 as well as their capacity to offer “informaciones precisas en la elección y empleo correcto de las palabras y locuciones que transmitan una idea dada, en un contexto lingüístico dado.” 23 Every dictionary must, then, help learners obtain and understand not only the meaning of words, i.e., the notional part of their description, which does not necessarily have a linguistic or scientific base, but also how those words function at the rest of the different levels of their characterization, such as orthographical, phonological, morphological, syntactic, and even pragmatic. Following Hornby (1954: v), “a knowledge of how to put words together is as important as, perhaps more important than, a knowledge of their meanings.” Furthermore, “modern lexicography considers dictionaries to be tools that are, or should be, conceived to assist a specific group of users in solving problems related to specific types of extra21 A badly designed bilingual dictionary, conceived as a simple repertoire of lexemes about which merely lexico-semantic information is given, is a double-edge blade and with no handle. A good bilingual dictionary, on the contrary, is a Swiss Army knife with multiple uses: it must help to decode or interpret, and to encode or produce in each of the two implied languages, informing in a precise way about collateral aspects of phonetics, syntax, pragmatics, register, and use; and trying to formalize or reveal the ideal and internalized knowledge existing in the mental lexicon of a native speaker (my own translation). 22 Fundamentally decoding instruments (my own translation). 23 Precise information on the choice and appropriate use of words and locutions which reflect a given idea, in a given linguistic context (my own translation). 21 lexicographic user situations,” these including text reception and production, translation, and the study of languages (Bergenholtz & Tarp 2006: 577). Be that as it may, not much information has been found about the arguments against the participation of grammar in dictionaries. The only one we have encountered refers to Spanish monolingual dictionaries and mainly revolves around the following ideas: (1) that they, i.e., monolingual dictionaries, generally have a decoding function, and (2) a presumed existing grammatical knowledge of the user (even though it may just be of an intuitive character) that makes this information redundant (Alvar Ezquerra 1993: 101-102). Those two points, though true to a certain extent, do not justify enough the exclusion of grammar in this type of works. On the one hand, no one can deny that dictionaries have an inherent decoding function, but it is also clear that, as mentioned above, they also have an encoding function that helps learners create syntagmatic structures of different sorts (phrases, simple sentences, or even complex sentences). This is precisely what leads us to a description of dictionaries as bi-functional tools. On the other hand, with respect to the presumed grammatical knowledge alluded in (2), only its existence may be presumed, since users are usually aware of certain aspects of grammar. How wide or complete that knowledge is or its level of internalization in the user’s mind, on the contrary, cannot be assured or even presumed. They, especially in their early stages of learning, may need to extend and/or consolidate their grammatical knowledge and the dictionary, as a pedagogical tool, should assist them on those matters. Besides, as Jackson (1985: 56) clearly points out: While the user can be expected to come to the Dictionary with a knowledge of the general rules of Grammar, the place of an individual lexical item within those rules (e.g. which of them apply to that item) is part of the idiosyncratic information about the item. All in all, may it be for decoding or encoding, extending or consolidating, learners can highly benefit from the participation of grammar in dictionaries during all stages of their learning process. This positive effect is the main reason why such information should not be disregarded or even treated lightly when elaborating any type of dictionary. 1.4.1. Grammatical information in dictionaries Having briefly discussed the relation between grammar and dictionary, we have concluded that the former should be always included in the latter. Now, the questions of how and to what extent that information should appear need to be answered. 22 “Dictionary making is tedious in the extreme. It is exacting. It is an incredibly large job” (Gleason 1967: 88). It requires great effort and time on the part of the lexicographers, who must constantly keep themselves updated regarding new words, new uses, and new nuances that may develop during any linguistic and communicative exchange. To this already existing and complex issue, the inclusion of grammatical information is added. One of the premises underlying lexicographic theory is that dictionaries must be simple, user friendly, but at the same time, precise, providing as much relevant and useful grammatical information related to the headword (also known as lemma, lexeme, or dictionary entry) as possible. In this sense, dictionaries should mirror a communicatively efficient language, which is shaped by a balance or trade-off between a pressure for simplicity and a pressure for informativeness or “expressivity” (Kirby et al 2015; Levshina 2022). Achieving the perfect equilibrium among these requirements is utterly challenging and, as such, not always successfully accomplished in actual lexicographic practice. We may encounter two different scenarios but with similar results. In an attempt to provide as much information as possible, a dictionary may end up being quite difficult to understand for the users, thus deviating from that sought simplicity and, in turn, from its primary purpose: serving as an aid to second and foreign language learners. And vice versa, while aiming towards that simplicity, some of the needs of the user may be overlooked, turning the dictionary into a useless tool for them. As the ones in charge of compiling dictionaries, lexicographers have to be very conscious of this balance. They, whilst bearing dictionary users and their needs in mind, have to make several decisions regarding the treatment of grammar that will squarely affect the usability of the dictionary itself. The components appearing in the microstructure of each dictionary may vary both quantitative and qualitatively speaking (Ramos Cortés 2006: 144). The information included as well as the method used to present it ultimately depends on the specific dictionary consulted. Among the great deal of grammatical information reflected in dictionaries, there are several kinds we normally expect to find: part of speech, references to inflections, syntactic behavior, government, and implicit grammatical information (Hoenigswald 1962: 108; Jackson 1985: 54; Mugdan 1989: 126). Let us now look at a brief description of the participation and importance of these components in dictionaries. It is commonplace to find the part of speech or word class as the first type of information that appears following the headword. This indication is extremely useful and relevant for learners because it allows them to locate a particular word within the grammatical system. In this way, it also provides them with further information related to word class: it gives them 23 information about the typical kinds of inflection that word may take (e.g., plurality in the case of nouns or past tense in the case of verbs) and about its possible syntactic behavior. This information has been refined over the years, distinguishing not only among the main word classes, namely noun, pronoun, verb, adjective, adverb, preposition, determiner, conjunction, and interjection, but also reflecting the fact that they are heterogeneous and that there may be subclasses (also named “secondary word classes” by Quirk et al (1985: 73)). That is to say, not all members of a certain part of speech take the same inflections nor behave in the same way —not all nouns admit plural or all verbs require an Object, for instance. Yet, in spite of that improvement, there are still various shortcomings, especially with respect to the distinction of verb typologies, an aspect that will be pointed out and further developed when dealing with the specification of the syntactic complementation. After the part-of-speech label, the inflections are usually found. Regular inflectional forms do not customarily appear in dictionary entries since they are either constant or regularly predictable and follow the general rule. Irregular inflectional forms, however, are not regularly deducible from the general set of grammatical rules (e.g., irregular plural forms such as shelves and teeth, or irregular past tense forms as saw) and do need to be explicitly indicated in the actual entries. Furthermore, in many cases the irregularity does not only imply a slight deviation from the general rule (smell – smelt), but a drastic orthographic change in the word itself that makes it almost impossible to discern the original lexeme (go – went). Not providing such information would inevitably lead learners to an error of overgeneralization, i.e., to the assumption that all words take the same regular inflections, as well as to the failure to recognize a particular inflection in a given context. This would definitely incite an incomplete and inaccurate competence that would translate into subsequent errors during the performance. As Jackson (1985: 55) states, “a learner needs maximally differentiated information in order to perform correctly and to be aware of available choices in the language.” Accordingly, to assure that irregular forms are easily found within the dictionary, apart from including them in the corresponding entry, dictionaries also add these “irregular and unexpected inflectional forms in the word list as reference lemmata that refer the users to the basic word form or entry, where all relevant data on the word is given” (Bergenholtz & Tarp 2006: 577). Something worth mentioning is that, among those consulted for this study, there are two dictionaries which also add regular inflectional forms in their entries. These are the COBUILD dictionary, which includes the complete verbal paradigm, and the Random House, which specifies the present and past participles of verbs as well as the comparative and superlative forms of one-syllable adjectives. 24 One of the most important kinds of information in dictionaries, together with meaning, is syntactic behavior. It is fundamental for learners to know how to accurately use lexical items in their own performance in order to be able to form and create grammatically correct and meaningful sentences: Learners of English as a foreign or second language, like learners of any language, have traditionally devoted themselves to mastering words — their pronunciation, forms, and meanings. However, if they wish to acquire active mastery of English, that is, if they wish to be able to express themselves fluently and accurately in speech and writing, they must learn to cope with the combination of words into phrases, sentences, and texts. (Benson et al 2010: xiii) As previously noted, a lexical item’s belonging to a specific word class does not entail the total sharing of the same properties and behavior, especially syntactic. Word classes are heterogeneous and within them several distinct subclasses can be distinguished, the most remarkable example being that of verbs. Verbal typology is extremely complex and varied. There are, among others, intransitive verbs, transitive verbs, and copulative verbs, and none of them requires or combines with the same elements, that is, none of them shares the same syntactic pattern (/SV/, /SVO/, and /SVC/, 24 respectively). Hence, the syntactic behavior cannot be deduced from the general rules of grammar but is an idiosyncratic property of the individual lexical items. This is precisely the reason why its appearance in dictionaries is so necessary and indispensable, because without this kind of information available, numerous ungrammatical sentences could be created. 25 A more detailed and in-depth analysis of this particular type of information in dictionaries will be made in a separate section (namely section 1.4.2.), given that it is one of the primary foci of this investigation. As stated by Hoenigswald (1962: 108), “government” 26 is one field of grammar whose appearance is particularly recurrent in dictionaries. It could be defined as the co-occurrence of lexical items with elements such as prepositions, affixes, and constructions outside the domain of inflection and derivation. This is also called “grammatical collocation” by Benson et al 24 These formulae stand for the syntactic constructions Subject + Verb, Subject + Verb + Object, and Subject + Verb + Complement. Syntactic formulae will be housed within slashes, following Garrudo Carabias (1991, 1996) and Comesaña Rincón et al (2013). The use of this typography “highlights the inherent abstract nature of those patterns, in parallel fashion to the symbolical procedure used to refer to another abstract term such as the phoneme, /p/, /t/, etc.” (Comesaña Rincón et al 2013: 38, footnote 16). 25 The term ‘ungrammatical’ refers not only to not following the grammatical rules of language, which would just be a syntactic analysis, but it also includes semantic and pragmatic examination. As it is well known, even though the famous sentence Colorless green ideas sleep furiously is syntactically grammatical, semantically and pragmatically speaking, it is completely nonsensical, as pointed out by Chomsky (1957: 15). 26 Mind this term as it is also used in Generative Grammar and the Binding Theory with a different use from the one understood in this study. 25 (2010: xix), who describe it as a phrase consisting of “a dominant word (noun, adjective, verb) and a preposition or grammatical structure such as an infinitive or clause” (e.g., decide on, be likely + to-infinitive, or be afraid + that-clause). Yet, a relevant observation should be made here. Prepositions are not the only grammatical words that can appear apart from grammatical constructions as Hoenigswald (1962: 108) and Benson et al (2010: xix) seem to suggest. We can also find other word classes like non-contrastive adverbs, i.e., those adverbs which do not allow free variation. This, for instance, would be the case of phrasal verbs. Another type of collocation that can be found in dictionaries is “lexical collocation.” This differs from the former type in the fact that in this case the two elements have equal status, i.e., there is no dominant word whatsoever, and that it “normally do[es] not contain prepositions, infinitives, or clauses. Typical lexical collocations consist of nouns, adjectives, verbs, and adverbs” (e.g., warm regards or set a record) (Benson et al 2010: xxxi). In dictionaries this sort of information is absolutely crucial if the large amount of phrasal, prepositional, and phrasalprepositional verbs is considered as well as its importance for second or foreign language learners as a basis to produce natural-sounding language. 27 All the different types of grammatical information mentioned above are explicitly incorporated in dictionaries. Nevertheless, that is not the only possible way to include grammar in them; it can also be implicit in the very selection of the entry through the use of illustrative examples. “They are particularly numerous and prominent in learners’ dictionaries, where they are seen as playing a crucial role” (Jackson 2013: 181). According to Cowie (1978: 129), illustrative examples have three main functions: (a) illustrate the syntactic behavior of the various senses that a lexical item may show; (b) clarify the meaning or different meanings of the lexical items; and (c) encourage learners to improve their performance and create lexically and syntactically original sentences. The advantages and benefits of including examples in dictionaries are universally accepted due to their multifaceted power forasmuch as these encapsulate three different areas: grammar, meaning, and use (Jackson 2013: 181). It is worth highlighting that this does not mean that the usual grammatical statement should be completely replaced, but that, if it is relatively minor (for instance, regular inflections or the ability of most adjectives and adverbs of appearing in both attributive and predicative position), an example could be used instead, considering that sometimes learners find them more clarifying than a proper explanation. 27 It is worth pointing out that this distinction between phrasal, prepositional, or phrasal-prepositional verbs is largely overlooked by dictionaries. Rather, the label “phrasal verb” is used to refer to all of these compound or complex verbs. See section 1.4.2. of the present study. 26 1.4.2. Syntactic and semantic aspects of complementation in dictionaries In order to get a full view of the behavior of different lexical items, there are two pivotal grammatical areas which should be considered when elaborating any lexical entry for a dictionary: syntax and semantics. 28 Apart from their own importance as separate subareas in Linguistics, they are also greatly interrelated and participate from one another, as is well known. As a matter of fact, it could be stated that the semantic content governs the syntactic functioning of the lexical item (Levin & Rappaport Hovav 1991: 124). In other words, there is a clear syntactico-semantic interface which plays a notable role within the realm of linguistic description. The significance of this relation has not been overlooked by various grammatical trends. To name some examples, the deep interpenetration and interdependency between syntax and semantics is what constitutes the basis of linguistic description in Functional Grammar, theory developed by Dik (1978) (Martín Mingorance 1990: 122), as well as one of the main hypotheses behind Cognitive Grammar (Cuenca & Hilferty 1999) and Valency Grammar (Herbst 1987). However, despite syntax and semantics being equally relevant and being intricately and closely related, their treatment in dictionaries is far from being the same. Referring back to the previous section, it can be noticed that whereas syntactic complementation has always received certain attention and study in lexicographical works, the inclusion of semantic roles is non-existent in almost every dictionary, if not every. In terms of syntactic complementation, it is necessary to emphasize that although there is an explicit reference to it in dictionaries, the aforementioned participation is not extensive enough and presents many shortcomings. Syntactic complementation is extremely complex, not only for the verbal category, which shows the greatest diversity and carries the information relative to intrasentential syntax, but also for the rest of word classes. 29 Moreover, it occupies a prominent part in the process of language learning, regardless of it being first, second, or foreign language learning. Garnica Silva (1991: ix) asserts that “un hablante es capaz de codificar o descodificar su pensamiento en una lengua determinada cuando es capaz de usar esa 28 At the syntactic level, a sentence is viewed as structure consisting of a verb and its completing elements. At the semantic level, a sentence is viewed as a communicative architecture which conveys a message and is focused on an element that predicates something. Hence, the term ‘semantics’ does not only refer to the meaning of a particular lexical item, but also to the semantic (or thematic) roles, i.e., the semantic relations between the different elements in a clause or sentence and the verbal unit. 29 Refer to Comesaña Rincón (1987, 1989) regarding the adjective, and Herbst (1988) and Tamayo Morillo (1998) with regard to the noun. 27 fructífera interacción entre léxico y sintaxis.” 30 That is to say, being able to encode and decode in a specific language does not just cover knowing the meaning of a certain number of words, knowing how they function in correlation with other lexical items is required as well. As already asserted by Nation (2001: 106), “knowing a word involves knowing how to use it in sentences.” Since this study deals with the verbal category (in particular with the physical perception verb class), we are going to focus our attention upon it. Verbal forms are not univocal and, consequently, they are likely to necessitate different arguments depending on the sense and/or the communicative intent conveyed, and, in turn, they show a vast variety of syntactic patterns. Lexicographers do conceive this possibility of a single verbal form having more than one completing behavior, yet when they put this thought into practice in dictionaries, they do not reflect it fully or accurately. In their attempt to present such diversity within one and a single verbal lexical entry whilst keeping the clarity and easy-to-use characteristic in their works, they usually end up making a vague and sometimes even misleading description of predicates by just distinguishing between intransitive, transitive, and copulative verbs (generally called ‘linking verbs’ in dictionaries). This means that out of the 25 different types of verbal patterns originally proposed by Hornby (1954), 31 they normally only take into account three of the main types, transitive, intransitive, and copulative, without considering the adjunctive one, let alone their corresponding subtypes. There are cases in which these also appear, but they are usually not well reflected in dictionaries as they are not even described as verb types. This situation seems rather restrictive, speculative, and, undoubtedly, far from a faithful description of the nature of verbal forms. Let us now focus our analysis upon the dictionaries consulted for this study. They are: Cambridge Advanced Learner’s Dictionary (McIntosh 2013), Collins COBUILD: Advanced Learner’s Dictionary (Sinclair & Combley 2023), Longman Dictionary of Contemporary English (Richards & Schmidt 2010), Macmillan English Dictionary for Advanced Learners (Rundell & Fox 2007), Oxford Advanced Learner’s Dictionary (Lea & Bradbery 2020), and The Random House Dictionary of the English Language (Flexner & Hauck 1987) (henceforth, CALD, COBUILD, LDOCE, MEDAL, OALD, and Random House, respectively). 30 A speaker is able to encode and decode his thought in a given language when he is able to use that fruitful interaction between lexicon and syntax (my own translation). 31 This author was the pioneer in attempting to provide a full treatment of the English verbal complementation patterns. These patterns are not only present in his theoretical work, but also in lexicographical ones, such as the Oxford Advanced Learner’s Dictionary of Current English (1974) and its following editions, e.g., the one edited by Crowther (1995). 28 These dictionaries seem to conform to the standard situation and MEDAL represents the exact image of it. Yet, there are slight variations as far as the rest of dictionaries are concerned as will be shown below. There are two opposite poles in terms of the quantity of syntactic information they offer: Random House and COBUILD. Random House stands at one of the extremes for it only makes a distinction between transitive and intransitive verbs, overlooking the usually present copulative type and providing, in our opinion, a very poor description of the verbal category. At the other extreme is COBUILD. It shows seven separate patterns, yet without any label associated to each of them nor any internal classification. In this way, it incorporates patterns for the main four verb types (intransitive, transitive, copulative, and adjunctive) as well as for three more subtypes of the transitive type, namely ditransitive, complex transitive, and adjunctive transitive. 32 There are three other dictionaries which also reflect the existence of the usually neglected main fourth type, i.e., adjunctive: CALD, LDOCE, and OALD. These, nevertheless, do not describe it as such. Both CALD and LDOCE include it in the form of a pattern or syntactic formula ([+ adv/prep] and [always + adv/prep], respectively) as opposed to the rest of verb types, which are represented by means of labels ([I], [T], and [L] or [linking verb]). The fact that the adjunctive typology is expressed through a different method creates a sort of division or separation between them as if they did not belong to the same level of characterization and it were more of a use similar to [+ infinitive] or [+ that] instead of a proper verb type. OALD also includes it using a similar formula ([v+adv./prep.]) and without an appropriate label. In this case, the whole verbal typology is expressed through the use of patterns or formulae, though this dictionary also adds a key to verb patterns in the front matter with a classification where this very pattern appears as a subtype within intransitive verbs rather than as a main type. Moreover, CALD and OALD distinguish three more subtypes each. By means of formulae, they signal the existence of ditransitive verbs —or as OALD labels them “transitive verbs with two objects”—, complex transitive verbs, and adjunctive transitive verbs, the latter accurately incorporated as a subtype within transitive verbs in OALD. As for their online versions, Cambridge, Longman, Macmillan, and Merriam-Webster (the online version of the Random House) remain exactly the same and use the same notational system. The remaining two dictionaries, on the other hand, present several modifications. Oxford also incorporates the main four typologies along with the subtypes but substitutes the 32 This typology is due to Comesaña Rincón (2000: 209-223). 29 aforementioned patterns for different ones, such as [somebody + adj.], [something + adv./prep.], or [+ adv./prep.], and uses the labels “intransitive,” “transitive,” and “linking verb.” This modification can be rooted in a seek for simplicity due to the fact that online dictionaries are accessed by a wider scope of users. With respect to Collins, it shows a shocking contrast with its printed version as it shifts from distinguishing between seven patterns to making no distinction relative to verbal typology at all. There is, however, a formula including the required elements after each example, even though it is not restricted to verbal typology. Table 1. outlines the data discussed above, i.e., the syntactic information included in the dictionaries consulted as well as the method used to present such information. The vertical axis comprises the type of dictionary and, thus, is divided into two, printed and online. The horizontal axis, on the other hand, represents verbal typology and is also divided into two, main types and subtypes of the transitive type. MAIN TYPES SUBTYPES Intrans. Trans. Copul. Adjunct. Ditrans. Adjunct. Trans. Compl. Trans. PRINTED CALD ✓ ✓ ✓ ✓ ✓ ✓ ✓ COBUILD ✓ ✓ ✓ ✓ ✓ ✓ ✓ LDOCE ✓ ✓ ✓ ✓ - - - MEDAL ✓ ✓ ✓ - - - - OALD ✓ ✓ ✓ ✓ (as subtype) ✓ (as main type) ✓ ✓ Random House ✓ ✓ - - - - - ONLINE Cambridge ✓ ✓ ✓ ✓ ✓ ✓ ✓ Collins - - - - - - - Longman ✓ ✓ ✓ ✓ - - - Macmillan ✓ ✓ ✓ - - - - Oxford ✓ ✓ ✓ ✓ ✓ ✓ ✓ Merriam-Webster ✓ ✓ - - - - - Label Pattern Table 1. Representation of the verbal typology in the dictionaries consulted. It has to be noted as well that labels are not always accurately used. To name a few examples, Merriam-Webster classifies the verb feel in I feel sick as intransitive when it is actually copulative, as it requires a Complement, and Cambridge (online version) classifies the 36 Online dictionaries function in the same way as printed ones, they usually list the most frequent meanings of a verbal unit without considering the rest. However, as much as space restrictions are a reasonable and solid explanation for this practice in printed dictionaries, it cannot be applied to the online format because of one of its main characteristics: its limitless nature. An online dictionary can list each and every meaning a verbal form conveys without it intervening in its manageability. In this case, the motivation for the exclusion or non-inclusion of these senses could be grounded on one or both of the following reasons: a) given that the rest of the senses are not that frequent, they are not considered significant and, thus, are excluded on the basis of simplicity; and/or b) online dictionaries may just be a mere reflection of the physical versions in terms of the number of lexical entries and senses they include. Be that as it may, as has been emphasized throughout this study, the lack of a full description of verbal forms may result in various misunderstandings and in an incomplete lexical acquisition, which can potentially stand as a hindrance to the performance of foreign or second language learners. Since online dictionaries are not affected by the aforementioned constraint, they are the perfect candidates to provide a full account of the verbal category and, in addition, they can be easily accessed everywhere, making the inclusion of a full description in them very convenient as well. According to Prototype Theory, 39 not all the meanings that a single verbal form can convey are equal nor hold the same status; quite the opposite, there are subtle hierarchical relations among them. Ibarretxe-Antuñano (1999) distinguishes between two types of meanings, those which are considered prototypical and those which are non-prototypical. The former type constitutes the basic or core meaning and is the most general and denotative. Customarily, it is the first one that appears in dictionary entries and the one chosen as the common answer to the question of what the meaning of a verb is. The latter, also called “extended meaning” by Levin and Rapoport (1988) or “peripheral meaning” in Prototype Theory, is described as “all those extended meanings, both physical and metaphorical, that these verbs can convey apart from the central prototypical meaning” (Ibarretxe-Antuñano 1999: 53). Through these definitions, it can be assumed that both kinds of meanings are related among 39 Prototype Theory posits that the internal structure of categories develops around a continuum or radial hierarchy, which is composed by a central element, the most representative or prototypical one, and other peripheral ones. These elements are related among them through what is known as family resemblance. This notion, originally proposed by Wittgenstein (1953) and later adopted in Prototype Theory, claims that the members of a category have at least one —but generally more— attribute in common with another member and not necessarily with all of them. Hence, “the members of categories which are considered most prototypical are those with most attributes in common with other members of the category and least attributes in common with other categories” (Rosch & Mervis 1975: 573). For further information, see Rosch (1973, 1975, 1978), Rosch and Mervis (1975), Cuenca and Hilferty (1999: 31-64), and Langacker (2006), among others. 37 them and that mentioned relationship is a hierarchical one, as claimed by Prototype Theory, with the prototypical meaning at the top and the non-prototypical meanings unfolding down at the bottom. By having a look at the labels used to refer to non-prototypical meanings, it can be noticed that whereas ‘extended meaning’ portrays a static picture with every meaning at the same level, ‘peripheral meaning’ denotes a gradable scale (that is why we can talk about a more peripheral meaning but not about a more extended one). Such a difference in the terminology clearly reflects one of the main hypotheses in Prototype Theory, the one which establishes that the hierarchy does not stop there, at the division between prototypical vs. non-prototypical meanings, but continues throughout the non-prototypical meanings in a way that the more peripheral a meaning is, the more semantic nuances it will show (Fernández Jaén 2005: 394, footnote 5). This is not exclusive to the various meanings a single verbal form can convey. That hierarchy of peripherality can also be applied to the different verbal units within a particular semantic field. Let us take, for instance, see, look, and observe. See shows the most general and denotative meaning within the semantic field of sight. Then, look would be more peripheral than see. In the same way, observe would be more peripheral than look and, obviously, see, for it has the most semantic nuances among the three. Non-prototypical meanings are thought to be derived from the prototypical or core meaning, hence they are also referred to as ‘extended.’ 40 The different linguistic environments in which a verb may appear affect its meaning by modulating it and this modulation, in turn, entails the creation of new meanings. 41 These new meanings may create new verbs that resemble the original verb inasmuch as they share the same form, but they may well differ not only in their syntactic functioning but also in the way the participants concerned participate in the communicative situation created. Regarding the core meaning, it could be stated that it has not been entirely lost or replaced by the new extended one, but that it is still present and has just been re-created or modulated, even though sometimes the modulation is so radical that it makes it difficult to discern that prototypical meaning. This means that verbs have a lexemic identity in different contexts, either linguistic or situational, that remains unalterable (Lyons 1977: 550-569). Therefore, we can establish several factors that can take part in this process of extension (Comesaña Rincón 2015): 40 There are linguists who are uncertain about this derivation process as well as about the existence of a core meaning. Nonetheless, this question falls beyond the scope of the present study and will not be discussed here. For further discussion, cf. Comesaña Rincón, J. & F.J. Tamayo Morillo (in preparation). 41 Prof. J. Comesaña Rincón and Prof. F.J. Tamayo Morillo, personal communication. 38 (a) The inherent lexemic core meaning. (b) The newly-created extended meanings. (c) The participants involved in the situation and their roles. (d) The likely syntactico-semantic process or processes involved (e.g., causativity, deletion, or metaphorization, among others). (e) The communicative intention and lexical co-occurrence. The inherent lexemic core meaning serves as the starting point for this process. Because of its inherent quality, it tends to be more stable and does not disappear over time. Thence, it normally provides the base, what may be modulated under certain linguistic environments or other factors to give rise to new meanings and verbs. The newly-created extended meanings are not as stable and can have a more sporadic or ephemeral use. In fact, as pointed out by Fernández Jaén (2012: 130), new meanings can be created ad hoc at a particular time, without it ensuring that they will be lexicalized in a permanent way. Extended meanings determine the syntactic and semantic complementation properties required in a specific context in the same way that the core meaning does. To put it differently, that relationship between the core meaning and the extended meanings does not imply that they have to share the same behavior and the same semantic configuration. Each of these new meanings has its own features and its own syntactico-semantic frames, which may differ greatly from the ones required by the core meaning. As a consequence, there seems to be a paradox within verb classes. Verbal forms are grouped together taking into consideration the features and patterns of this prototypical meaning, but these do not always match those of the nonprototypical ones. Let us compare the following sets of sentences: (2) a. He broke the vase. b. My phone broke. c. He finally broke the code and opened the door. (3) a. I saw the man coming towards me. b. I can see perfectly. c. I see what you mean. As can be appreciated through these examples, a verbal form that represents verbs whose prototypical meaning belongs to the class of change-of-state verbs such as break, which is a pure example of this verb class, may behave in a similar or the same way as a cognitive verb when it conveys the meaning of ‘to decipher’ as in (2c). Alternatively, a physical perception verb such as see, the quintessential representative of this verb class, may behave as a cognitive 39 verb as in (3c) when it means ‘to understand.’ It can be concluded, then, that verbal forms do not compose homogeneous semantic classes for they are not homogeneous themselves. It is impossible to assure that a certain verbal form syntactically behaves in a specific way and necessitates specific arguments in each and every linguistic context. Going back to the previous examples, although the verbal form break usually has an Agent in the Subject position, in (2b) the Subject performs the semantic role of Affected and in (2c) of Agent-experiencer. Similarly, the verbal form see can be described as a transitive verb ((3a) and (3c)), but that would imply rejecting all those instances in which it is used intransitively, as in (3b), I can see (‘I have the capacity to see’) or I see (‘I understand’). 42 Such variations in the participants involved in the situation as well as in their roles is common, not to say possible, when new meanings are created. As mentioned above, the different meanings a single verbal form may convey are in charge of determining their own syntactico-semantic complementation properties. Some may coincide totally or partially syntactically and/or semantically, whereas others can show radically disparate complementation patterns. Metaphorization and the different syntactico-semantic processes that may be involved can be defined as the most significant devices to semantically extend the core meaning. Due to this and to the fact that it does not only deal with semantics but also syntax, this matter will be commented on in a separate chapter, namely Chapter 4. In the process of semantic extension, there is an unquestionable interaction with the pragmatic function. Verbs —and potential predicative units in general— are not entities that can be analyzed in complete isolation, as they are mostly used in conjunction with other lexical items and rarely alone (mind, for instance, some uses like fire!, fetch!, or out!). Depending on the communicative intention, the speaker chooses particular lexical items to co-occur with the main predicative unit of the sentence, i.e., the verb. The semantic properties of these items must be licensed by the verb and the other way round, in a pure communicative and grammatical synergy. This unit has the capacity to combine itself with multiple elements, but not all meanings that may be conveyed by a single verbal form can appear in any and all contexts. It is, indeed, the communicative intent, the linguistic environment, or the situational context that will influence or modulate its meaning. Let us compare the ensuing examples: 42 Actually, although apparently intransitive, the latter example is a case of pseudo-intransitivity. That is, in this example, see requires an Object for its syntactico-semantic completion (e.g., what you are saying or what you mean), but this element has been contextually elided. Hence, see is not inherently intransitive here as it has an implicit Object. 40 (4) a. He smelt the coffee. b. He smelt trouble. In (4) both the communicative intention and the elements that combine with smell are sharply different. As a result, we are dealing with two separated meanings, the one in (4b) created through the modulation of the prototypical one in (4a). These meanings cannot be expressed by combining smell with any lexical items, because they each have their own syntactic and semantic requirements. In the case of the meaning conveyed in (4b), it shows more restrictions since it is an extended or non-prototypical meaning. Hence, its combinatory possibilities as well as the situational and communicative contexts in which it can appear are narrower. In sum, the linguistic and communicative context in which verbs appear have a high influence in the final meaning conveyed, this being the very reason why, in order to express exactly what we want to say, we make use of specific words and constructions and disregard others. 2.2. Semantic classification of physical perception verbs When dealing with the analysis of any verb class and, by extension, of any grammatical category, some sort of internal classification is needed, otherwise it may result in a chaotic attempt to shed some light on an already complicated and intricate matter. A verb class cannot be analyzed as a whole, i.e., as one single entity with certain specific characteristics which are consistent throughout all of its members. In the same fashion, analyzing it as a group composed by several isolated and distinct verbs (see1, see2, see3, seen, seen+1), each showing particular features, will be found equally as misleading and untenable. The main reason behind this is that verb classes are not completely homogeneous per se, yet the verbal units that compose them are not completely discrete either. On the contrary, they do show some type of resemblance, whether it is semantic, syntactic, and/or structural, similar to the notion of family resemblance in Prototype Theory. As the meaning —the notional part of the verb— is the pivotal aspect of this predicative unit and semantics is closely related to syntax (refer to section 1.4.2. above), the method used in this investigation to internally classify the physical perception verb class is through a semantic classification. Physical perception verbs will be divided into six semantic fields and three parameters, 43 which results in eighteen different perceptual situations. 43 As will become clear, we understand the term ‘parameter’ along the lines of Garrudo Carabias (1999) and without bearing any relation with the way Chomsky (1981) uses it. 41 Regarding the semantic fields, physical perception verbs are generally distributed among five different fields which stand for each one of the senses human beings have, namely sight, hearing, smell, taste, and touch. This is the traditional classification made by most linguists, such as Viberg (1984, 2008), Ibarretxe-Antuñano (1999, 2002, 2019), and Fernández Jaén (2005, 2012), among many others. However, this largely common practice is not without problems. During the process of analyzing this verb class, we encountered a significant number of verbs that did not entirely fit into any of the five fields. In such instances, though it was clear that the verbs denoted perception, there was no specification whatsoever of the exact sense employed. The linguistic context did, in some cases, reduce the possibilities, but still it did not clarify exactly which specific sensory modality participated in the process or act of perception. 44 These verbs are usually called multimodal (Fernández Jaén 2012: 396) or multi-sense verbs (San Roque et al 2015: 40-41; Norcliffe & Majid 2024: 85). Thereby, in the present study a sixth field labeled general will be added following Garrudo Carabias’ (1999) work. Using his own words (1999: 181), “el último campo mencionado se refiere a la posibilidad de expresar una percepción física genérica de tipo global.” 45 That is to say, this last field includes those instances in which all or some of the sensory modalities are neutralized into just one verb. This can be illustrated in the example The dead cannot feel, in (5f) below, in which the verb feel does not only entail the sense of touch, but also the rest of them, since the dead cannot either see, hear, smell, or taste. Each of these six fields comprises various verbs, the most representative or prototypical ones being see, hear, smell, taste, touch, and feel, respectively. Apart from this sensorial distinction, physical perception verbs can be further subclassified into three parameters depending on the phenomenon called base selection, that is to say, the semantic role of the Subject required by the verb (Viberg 1984: 124). These parameters have received different labels by different authors throughout time, as will be noted in each one of their descriptions. In the present study, they will be referred to as PROCESS, ACTION/PROCESS, and STATE (Garrudo Carabias 1999, 2015): (a) Those physical perception verbs whose Subject-referent has the semantic role of Experiencer appear in the PROCESS parameter, which, if we stick to the traditional definition by Poutsma (1926: 341), is described as “the receiving of an expression by the senses 44 This situation is not exclusive of the verbal category. Adjectives and nouns pertaining to the perceptual domain have also been found to be capable of referring to more than one sensory modality at once (Lynott & Connell 2009, 2013; Speed & Majid 2017). 45 The last field mentioned refers to the possibility to express a generic physical perception of a global type (my own translation). 42 independently of the will of the person concerned.” 46 The verbs under this parameter present their Subjects as passively involved in the process of the perception or the “receiving” of a Stimulus, which is, if there is a need to mention it, in its turn, developed by the syntactic function of Object (mind (5f) whose Object anything has been omitted due to the loss of informative value or “communicative dynamism” as Quirk et al (1985: 1356) put it). Therefore, in this set of verbs there is no volition or intention involved in the process of perception on the part of the Subject-referent. This is exemplified in (5). 47 (5) a. Eric saw a bird. b. We heard a cry. c. I am unable to smell anything. d. Sam tasted garlic in the stew. e. She felt the breeze on her face. f. The dead cannot feel. (Garrudo Carabias 1999: 181) (b) The ACTION/PROCESS parameter covers those verbs whose Subject-referent acts as an Agent-experiencer. 48 This contrasts with the previous parameter in that the Subjectreferent holds a dual semantic role and is no longer described as a mere Experiencer, but also as an Agent. 49 Such agentive role inevitably brings about a new implication, the willingness or volition on the part of the Subject-referent (obligatorily an animate participant unless metaphorization is involved) in the process or, in this case, also action (hence the name) described by the predicative unit. As in the PROCESS parameter, there is a Stimulus which, if 46 Other terms that have been used to name this set of verbs throughout the linguistic literature are “cognition” (Rogers 1971: 206; 1972: 304), “experience” (Viberg 1984: 123; 2008: 124; Ibarretxe-Antuñano 1999: 45; Huddleston & Pullum 2012: 169), “inert perception” (Leech 2004: 28), “private” (Palmer 1988: 72-73), or “passive” (Blendea 2015: 92). 47 The verbs used in the examples of each of the parameters are considered the most representative or prototypical and do not constitute the whole group of verbal units that will be examined here. 48 This parameter has also received several names, some being “active perception” (Rogers 1971: 206; 1972: 304; Leech 2004: 28), “activity” (Viberg 1984: 123; 2008: 124; Ibarretxe-Antuñano 1999: 45), “action” (Garrudo Carabias 1999: 182), “acquisition” (Huddleston & Pullum 2012: 169), or “active” (Blendea 2015: 92). 49 This coalescence of more than one semantic role for the same argument goes against the characterization of thematic roles shared by Chomsky (1981), Bresnan (1982), Carlson and Tanenhaus (1988), or Dowty (1991), among others. According to them, one of the features of the θ-criterion of the Government and Binding theory is that “[e]ach argument bears one and only one θ-role” (Chomsky 1981: 36, 139, endnote 14). However, quoting Mora Gutiérrez (2001), “it is doubtful whether [this view] can be maintained. Uniqueness [sic] does not seem to hold with animate subjects of verbs of motion.” As linguists such as Gruber (1965: 216-217) and Jackendoff (1972: 32) have already pointed out, the Subject in these cases “is both agent, since it initiates and sustains the movement, and theme, since it is the object that moves” (Mora Gutiérrez 2001). 43 overtly mentioned, is syntactically developed as the Object, with the exception of look, which requires an Adverbial instead: 50 (6) a. Eric looked at the bird. b. We listened to the radio. c. Smell that flower. 51 d. Sam tasted the stew. e. She touched the cloth. f. Notice the way I do it. In the light of these two parameters, Gisborne (1996: 13, footnote 5) proposes the socalled “DELIBERATELY [sic] test” in order to distinguish between them. He claims that any verb that can co-occur with manner adjuncts like deliberately or with a purpose clause belongs to the ACTION/PROCESS parameter, because they signal the “agency” (in Gisborne’s (1996: 13, footnote 5; 1998: 14) terms) of the Subject-referent which is characteristic of this group. If, on the contrary, a verb does not accept such adjuncts, then it will be automatically included in the PROCESS parameter, since it will suggest the unwillingness and lack of volition of the Subject-referent. This can be illustrated in (7): (7) a. Eric deliberately looked at the bird. b. *Eric deliberately saw a bird. c. She listened to the music (in order) to check if it was good. d. *She heard the music (in order) to check if it was good. The similitude between the two labels, PROCESS and ACTION/PROCESS, does not indicate that the meanings of the verbs under the former parameter are necessarily implied in the latter. To put it differently, the fact that someone is looking does not mean that they can see something. 52 Thus, sentences such as I was looking carefully but couldn´t see a thing or I’m listening but can hear nothing are completely grammatical and, what is more, logical. Besides, another dissimilarity that is shown through these sentences is that the ACTION/PROCESS 50 As Quirk et al (1985: 90-91, 1155-1156) already signaled, this verb may also be analyzed as prepositional monotrasitive, subcategorizing a Prepositional Object headed by at. In the present study, however, this preposition is viewed as being in paradigmatic relationship with others, such as out in he looked out the window or up in she looked up the sky. Hence, this verb is here described as simple adjunctive. 51 Mind the sentence I smelled the flower. This would be ambiguous as the Subject-referent can be understood as a passive or active participant, i.e., as showing volition or not, depending on the situational and communicative context. 52 It is worth mentioning that this statement would not hold were the verb look at rather than look. The reason is that look at inherently implies to see, that is, there is semantic inclusion between these two verbs. 44 verbs are considered to be non-stative or dynamic (following Lakoff’s (1966: 7) or Quirk et al’s (1985: 178) terminology, respectively), i.e., they admit the progressive and imperative constructions, whereas PROCESS verbs are classified as stative, i.e., they reject such constructions (see Chapter 3. for more information about non-stative vs stative verbs). (c) Lastly, the physical perception verbs whose Subject-referent is a Stimulus compose the STATE parameter. Compared to the previous ones, this is the most different in terms of both syntactic and semantic framings. In this case, the participant playing the role of Experiencer is not the grammatical Subject nor is it normally overtly mentioned in the sentence, but it is assumed to be the one making the assessment unless otherwise stated (see below). In fact, taking example (8a) below as a reference, the person who looks pretty may be blind, what makes it clear that the Subject you is not the Experiencer of look. In other words, you is not an argument of look, but of pretty. 53 One more difference is that this group of verbs requires a Subject Complement with the semantic role of Attribute, more specifically, Current Attribute, instead of an Object: 54 (8) a. You look pretty. b. You sound sad. c. The cheese smells good. d. The wine tastes delicious. e. This silk feels soft. f. It feels good to be happy. These major differences in the semantic characterization of their Subjects between the previous two parameters and this one have led to a higher distinction that groups them into experiencer-based and phenomenon-based (Viberg 1984: 124). The former class comprises the first two parameters, PROCESS and ACTION/PROCESS, as they both involve an Experiencer 53 Actually, the verbs under this parameter are not copulative verbs strictly speaking. They are instances of raising verbs, which can be described as “[v]erbs like seem which take a raised complement” (Huddleston & Pullum 2012: 1194). That is to say, these are verbs whose syntactic Subject does not correspond with the semantic one; rather, it is an argument of an embedded clause, the Subject, which has been raised to the Subject position of the primary predicate, look. On the basis of this description, we may have an alternative analysis: (i) Youi look ti pretty → [you are pretty] looks This analysis also affects the semantic characterization, i.e., the thematic roles. Given the two predicative units (look from the main clause and pretty from the embedded one), there is a double predication resulting in a dual semantic role of the Subject, namely Stimulus-affected. 54 Some other labels used to refer to this parameter are: “flip verbs” (Rogers 1971: 206; 1972: 304), “copulative” (Viberg 1984: 123; Blendea 2015: 92), “verbs of state” (Palmer 1988: 71), “percept” (Gisborne 1996: 1; Ibarretxe-Antuñano 1999: 45), “phenomenon-based” (Viberg 2008: 124), or “production” (Huddleston & Pullum 2012: 169). 45 as their Subjects (with the addition of an agentive role in the second parameter as well). The latter class includes the last parameter, STATE, as the Subject of these verbs plays the role of Stimulus or Phenomenon, using Viberg’s (2008: 124; 2019: 18) terms. A different aspect that also sets this parameter apart from the rest is that the verbs belonging to it are epistemic, i.e., they express subjective information from the Experiencer’s own knowledge or perception of reality (Huddleston & Pullum 2012: 178). This can be proven by adding a to-phrase, whose Prepositional Object would stand as the Experiencer, at the end of each of the examples in (8) without it altering their meanings. 55 In most cases, this element is treated as optional and is only expressed in those situations in which the subjective nuance of the statement wants to be emphasized or when the Experiencer is not the speaker, but someone else. In Gisborne’s (1998: 11) words: I take it that the TO-phrase data is evidence of subjectivity and that a subjective interpretation of these verbs is further evidence that they are epistemically modal. […] By default, that experiencer is the speaker. Otherwise, the experiencer is referred to in the TO-phrase. Having laid out the semantic fields and the parameters, the semantic classification of physical perception verbs along the lines of the coordinates established by Garrudo Carabias (1999: 185; 2015) with a few personal modifications would be as follows: 56 55 Just in passing, this to-phrase is referred to as datiuus iudicantis in classical studies or ‘dative of point of view’. Quoting Ernout and Thomas (1972: 72): “Datif de point de vue (datiuus iudicantis) indiquant aux yeux de qui l’affirmation énoncée est vraie: Catul. 86, I: Quintia formosa est multis «Quintie est belle aux yeux de beaucoup»” [Dative of point of view (datiuus iudicantis) indicates according to whom the affirmation stated is true: Catul. 86, I: Quintia formosa est multis «Quintia is beautiful according to many» (my own translation)]. 56 These personal modifications correspond to the incorporation of the coordinates “experiencer-based” and “phenomenon-based,” the specification of the syntactico-semantic complementation properties as patterns per se, and finally, the addition of the verbs stare, sniff, and savor. 52 CONTROLLED INSTRUMENT [CntrlInstr] Participant which is consciously used or handled by an Agent and which can be re-used indefinitely. CURRENT ATTRIBUTE [Curr] The attribution ascribed to the referent exists in the real world regardless the predication expressed. DESTINATION LOCATIVE [DestLoc] Participant which refers to a spatial location as an intended or predetermined end point (or destination). DESTINATION TIME [DestTime] Participant which refers to the temporal end point. EFFECTED [Eff] Physical participant that comes into existence as the result of the action or process denoted by the verb. EVENTIVE [Evnt] Participant which is not a physical entity or substance and is seen as limited in time, i.e., having a starting point and an end point. EXPERIENCER [Exp] Participant which experiences either physically, cognitively, or emotionally, and reacts to a Stimulus. FORCE [Frc] Participant that acts irresponsibly, unwilfully, and/or out of control and causes a process or change of state. INSTIGATOR [Instg] Participant that acts on an Agent and instigates his action. LOCATED [Lctd] Participant whose location in time or space is predicated by the verbal unit. MANNER [Mnnr] Participant that limits the way in which the verbal meaning is expressed, i.e., an action is performed or a process takes place. MEASURE [Measr] Participant expressing the objective value of substances through their quantification by showing their duration, length, money, time, width, etc. ORIGIN LOCATIVE [OrigLoc] Participant which refers to a spatial location of provenance or source. ORIGIN TIME [OrigTime] Participant which refers to the temporal location from which a state, process, or action departs. POSITION LOCATIVE [PstnLoc] Participant which refers to the spatial location as an end point, whether final, current, or stable. POSITION TIME [PstnTime] Participant which refers to the temporal location as an end point, whether final, current, or stable. REASON [Rsn] Participant which clearly states or describes the cause of the state, process, or action expressed by the predicative unit. RECIPIENT [Rcpt] Participant that receives something in a transaction or transfer. It can be considered as a temporary step, not the permanent or final aim. RESULTING ATTRIBUTE [Rslt] The attribution ascribed to the referent is triggered by the appropriate predicative unit. 53 STIMULUS [Stm] Participant involved in an emotional, sensory, or cognitive experience. TOPIC [Tpc] Participant that appears with predicates pertaining to metalanguage and communication fields. TRANSIT LOCATIVE [TransLoc] Participant which refers to a spatial location considered as a passage through which persons or things move. TRANSIT TIME [TransTime] Participant which refers to a temporal location or span considered as a passage through which persons or things go. Table 3. Semantic roles. 2.3.1. Two arguments In those instances in which a physical perception verb requires two arguments, the one that tends to show a greater variation is the first one, that is, the Subject. Such a fact is precisely reflected in the typical semantic frames of the three aforementioned parameters, as this participant can be an Experiencer, an Agent-experiencer, or a Stimulus. Nevertheless, these are not the only possibilities. They just correspond to the most common types of characterization. Depending on the specific meaning conveyed by the verb, the Subject-referent can express other semantic roles which may be far from the ones normally related to the physical perception verb class. As will be seen later on (section 4.3.1.), the verb see can have the meaning of ‘to witness.’ Syntactically speaking, it behaves monotransitively, showing the pattern /SVO/. Regarding its semantic roles, there are some aspects worth discussing. The Subject takes the role of Experiencer, as in (15a) below, thus the verb falls into the PROCESS parameter. However, the same role cannot apparently be applied to any of the rest of Subjects in the examples (15b-e): (15) a. I hope I never live to see the day when computers finally replace books. b. This stadium has seen many thrilling football games. c. This summer has seen the end of water restrictions. d. This pen has seen a lot of good writing. e. His hair looks like it’s never seen a brush! 62 For a participant to be an Experiencer, it needs to be able of experiencing, which means that it has to be animate, a characteristic that cannot be easily attributed to the referent of the head noun stadium or pen, for instance. The meaning of the verb see does not change throughout 62 The Object a brush needs to be taken metaphorically, i.e., metonymically: the instrument for the action. 54 any of these examples, we are indeed dealing with the same verb that requires an Experiencer acting as the Subject. How is this possible, then? Why does the verb license those NPs as its Subjects? The reason behind it resides in the phenomenon of metaphorization, which alters our vision of the extralinguistic referents. Through it, a place, a temporal space, an instrument, or an inalienable body part may be able to act as the Experiencer. Therefore, the Subjects in (15be) have to be taken metaphorically. For instance, This stadium has to be understood not as the place, but as the people in that place. As a consequence, these Subjects are, respectively, an Experiencer-position locative, an Experiencer-position time, an Experiencer-instrument, and an Experiencer-affected. There is a coalescence between two semantic roles, the one originally licensed by the verb and obtained through metaphorization, and the one reflecting the original identity of the Subject-referent. That is to say, there is a communicative synergy, an interaction between the meaning of the verb and the semantic nature of the NPs in this case. In this same parameter but in the general field, there is one more instance in which the Subject plays a non-prototypical role, so to say. This case occurs when the verbal form sense means ‘to detect,’ behaving monotransitively. The Subject is not an Experiencer, as could be expected, but a Controlled Instrument. The rationale for that characterization is similar to the one above, namely the interaction between the meaning of the verb and the semantic nature of the Subject-referent. Besides, the Object is not a Stimulus either, but an Affected participant: (16) a. A motion detector can sense movement. b. Equipment that senses the presence of toxic gases. In the semantic field of touch, there are some examples of verbs following a /SVO/ pattern and showing variation in their semantic roles. The verb touch can convey the meaning of ‘to border, to be in contact with,’ but its Subject and Object are not, once again, an Experiencer or a Stimulus. On the contrary, with this specific meaning, the Subject and the Object stand for participants that play different roles which do not follow the traditional description. (17) a. Our properties touch each other. b. Their hands were touching each other. In this case, the Subject plays the role of a Located-position locative and the Object the role of a Position Locative-located. Neither an Experiencer nor a Stimulus are expressed by any of the participants, showing a drastic deviation from the prototypical characterization of the physical perception verb class. The reason for the conflation of the same semantic roles in both 55 elements is due to the symmetrical nature of the meaning of the verb and, hence, the symmetrical relationship of its arguments. 63 Likewise, when the verb touch means ‘to touch or brush against something,’ the Subject is expressed by a Force in those cases in which it is an inanimate entity or an entity with no volition. Consequently, the Object is an Affected participant rather than a Stimulus: (18) Her dress was so long that it was touching the ground. There are some cases in the PROCESS parameter of the general field which require special attention for they could be considered peculiar or unique. These cases involve three different syntactic patterns, all of which are composed of two subcategorized units. The patterns in question are /SVC/, /SVPC/, and /SVA/, as respectively appear in (19): (19) a. I felt a complete idiot. b. I felt like a murderer. c. He feels as if he had been betrayed. Semantically speaking, none of these patterns has a corresponding Stimulus as the second argument; instead, they respectively require a Current Attribute in the first two and a Manner participant in the last one. Although this represents a deviation from the prototypical realization, this is not the aspect considered peculiar. Since the syntactic description also differs from the prototypical one, i.e., /SVO/, it is not surprising that the semantic characterization is affected in some way as well. What distinguishes these three verbs from the rest and makes them peculiar is that the Subject bears the semantic roles of Experiencer and Stimulus at the same time. At first glance, this may seem paradoxical, due to the fact that the Experiencer is the participant which experiences or perceives a Stimulus, which might imply that the two roles cannot or should not be associated with the same constituent in a sentence. Nevertheless, they do not occur in complementary distribution. In these very cases, the Subject-referents are perceiving themselves in a particular way. If it were otherwise, i.e., if the Subject did not play both roles, 63 Symmetry is a built-in component, an inherent feature of many predicative units which entails a logically reciprocal relation between the participants involved in the situation described by the verb. This reciprocity is also translated into the syntactic level, since those participants allow their reversibility or permutability as well as to be associated in the subjective slot: Eric met Nell = Nell met Eric = Eric and Nell met. As Comesaña Rincón et al (2013: 142-143) point out, lexical entries —and, hence, verbs— can be interconnected via symmetry either negatively or positively. It is said to be negatively interconnected when there is deletion of the symmetrical proform each other, for it lacks communicative relevance and “is felt to be situationally redundant” (Comesaña Rincón et al 2013: 143), as in (24) below. Conversely, it is positively interconnected when there is insertion of that symmetrical proform, which makes the verb “acquire a symmetrical quality out of a non-symmetrical one” (Comesaña Rincón et al 2013: 143). 56 these verbs would have to subcategorize one more element, specifically an Object developed by a reflexive pronoun, becoming complex transitive or adjunctive transitive verbs: 64 (20) a. ?I felt myself a complete idiot. b. ?I felt myself like a murderer. c. ?He feels himself as if he had been betrayed. In the ACTION/PROCESS parameter, there are several cases of monotransitive verbs which deviate from the prototypical characterization, requiring an Agent as its Subject and an Affected as its Object. This is the case, for instance, of look with the meaning ‘to plan,’ observe with the meaning ‘to obey,’ sound with the meaning ‘to pronounce,’ or touch with the meaning of ‘to partake of food or drink’: (21) a. My parents are looking to buy a new car early next year. b. You can avoid danger by observing these simple rules. c. You have to sound each letter. d. My grandfather doesn’t touch alcohol. With respect to the STATE parameter, there is little variation in the semantic roles that can be reflected on the syntactic elements proper. The Subject typically expresses the role of Stimulus, although in some specific cases it can be expressed by an Experiencer-stimulus, as happened in examples in (19). That is to say, it is possible that the Experiencer is perceiving themselves and the Stimulus, the Experiencer, and the speaker conflate into the Subject element. For this to be feasible, the Subject needs to be realized by a first-person pronoun, either singular or plural: (22) a. I sounded just like my mother. b. We look great in that picture. However, that implication is not necessarily present in every case in which such a condition is met. The sentences in (22) can also be understood as I sounded just like my mother [to other people] and we look great in that picture [to other people], for instance. Here, the Experiencer is not the speaker, but another participant, and, as a result, the Subjects would be mere Stimuli. The second argument can also show two alternative semantic roles. On the one hand, when the verb behaves copulatively, the Complement is an Attribute, more specifically a 64 This would correspond to the pronoun ‘-se’ in the Spanish verb ‘sentirse.’ 57 Current Attribute, following the prototypical frame. And on the other hand, if the verb behaves adjunctively, the Adverbial plays the role of Manner. The semantic role of the Experiencer, though not overtly expressed, is present in the figure of the speaker, given the epistemic feature of the verbs within this parameter. (23) a. That melody sounds terrifying [to me]. b. Her perfume smelt too sweet [to me]. c. Eric looks as if he had been crying [to me]. 65 2.3.2. One argument Physical perception verbs can also behave intransitively by requiring just one subcategorized unit, the Subject. In these cases, the Subject-referent normally expresses one of the three typical semantic roles in this position (Experiencer, Agent-experiencer, or Stimulus) depending on the parameter to which the verb belongs (see Table 2.). However, that does not mean that it is impossible for the Subject to bear a role different from the typical ones. There is an occurrence in the touch field in which the verb touch functions as an intransitive verb and its Subject is not expressed by the prototypical roles. The specific meaning conveyed by touch in this case is one which has been already discussed, but when behaving monotransitively, namely ‘to border, to be in contact with.’ Although apparently paradoxical, the intransitive counterpart has the same semantic characterization, [Lctd-pstnloc, PstnLoclctd]. That is to say, whereas the syntactic framing is monadic, the semantic one is dyadic. This is feasible due to the symmetrical nature of the meaning expressed by this verb. In fact, the only difference between (17a) and (24b) is that the symmetrical proform each other has been deleted in the latter, as it is not communicatively necessary or relevant due to that inherent symmetrical nature. Thus, given that this verb is inherently reciprocal, syntactically, the Subject is seen as encapsulating multiple referents, what explains the conflation or multiplicity of semantic roles that the Subject may take. Semantically, however, this verb is described as requiring two separate logical arguments: (24) a. Make sure the wires don’t touch. b. Our properties touch. c. They were standing side-by-side with their shoulders touching. 65 This is, once again, an instance of datiuus iudicantis (Ernout and Thomas 1972: 72). See note 55. 58 The verb smell of the STATE parameter can also be used intransitively when meaning ‘to smell badly.’ In a similar fashion to the verbs in (22), when the Subject is realized by a firstperson pronoun, it may have the dual semantic role of Experiencer-stimulus. (25) Don’t come closer, I smell. Last but not least, the verb sound can behave intransitively with the meaning ‘to produce a sound’ and ‘to pronounce.’ In both cases the Subject is realized by an Affected participant, deviating completely from the prototypical characterizations: (26) a. If the alarm sounds, leave the building immediately. b. English is not spelt as it sounds. 2.3.3. More than two arguments These verbs may also require more than two subcategorized units. The first two constituents are in most cases a Subject and an Object, which are normally developed by an Experiencer or an Agent-experiencer and a Stimulus, respectively (mind (27c), which requires an Agent and an Affected as its Subject and Object). The third element may be related to the verb in different functional ways. It may be an Object Complement, an Adverbial, or a Prepositional Object. Semantically, its characterization is also wider. For instance, some of the possible realizations are Current Attribute (27a), Manner (27b), Destination Locative (27c), Topic (27d), or Eventive (27e): (27) a. Bill perceived his comment as an attack. b. You have to look at it in a different way. c. Let me see you to the door. d. I have to see my teacher about my grades. e. The unemployment data must be seen against the background of world recession. 2.4. The semantic domain of evidentiality There is a semantic domain that should not be overlooked when dealing with physical perception verbs for it yields insights into their nature and structure: evidentiality. Before embarking on an examination on how this domain is present in the physical perception verb class, however, the actual notion of evidentiality must be explored. According to Jacobsen (1986: 3-4), this term was originally introduced into the realm of Linguistics about eighty years ago in a posthumously published grammar of Kwakiutl by Franz Boas (1947: 237, 245). 59 Evidentiality is a semantic-functional category that can be defined as the marking or linguistic encoding of the source of the information or evidence (hence the name) of a proposition (de Haan 2001: 201; Roseano et al 2016: 136). In other words, it signals whether the speaker actually witnessed what is described in the statement, heard it, knew about it through another person, or simply made inferences based on some evidence. The terminology used to refer to this notion is quite varied. For instance, Johanson (1996, 1998, 2000) uses the label indirective when describing evidentials in Turkic languages (e.g., in Turkish and Bulgarian), Adelaar (1977) —as well as other scholars who follow him— employs verificational or validational in his analysis of Quechua, and grammarians of Aymara such as Hardman (1986) use data source markers. 66 French linguists, on the other hand, normally prefer using the term médiatif or mediative (e.g., Guentchéva 1994, 1996) instead of the calque évidentiel in the light of the possible conceptual confusions it may bring about (Dendale & Tasmowski 1994: 3; 2001: 340; Guentchéva 1996: 13; 2004: 22-23, footnote 8). Those confusions are based on the fact that the terms evidential and évidentiel are false friends. The English term evidentiality derives from the noun evidence, but its meaning, ‘proof,’ differs from the one of the French noun évidence, ‘clear and manifest certainty that cannot be doubted,’ ‘something evident.’ Paying attention to its etymology, évidence comes from the Latin term evidens, evidentis, formed by e- (ex-), ‘out,’ and videre, ‘to see,’ which means ‘obvious to the eye.’ 67 In this way, whereas the English terms evidentiality or evidential show a wider semantic scope, encompassing all different types of ‘proofs’ or sources for the information, the semantic scope of the French calques évidentialité and évidentiel are narrower, excluding those sources that do not come from our visual perception and, thus, they are not considered appropriate or transparent enough to refer to this notion. 68 In spite of that, the term ‘evidential’ —and its calques in other languages— is still the most frequent in the linguistic literature across languages, probably in an attempt to favor a common ground for the crosslinguistic identification of this phenomenon. Whenever we communicate something, either orally or in writing, we are conveying some information which has previously been acquired through diverse means or, in more exact terms, 66 For further information, see Aikhenvald (2004: 15). 67 It needs to be noted, however, that although the rough translation of the Latin word refers to visual perception, evidens, evidentis shows three different meanings, as recorded in the Oxford Latin Dictionary, which are less restrictive in that sense: 1. Perceptible (to the senses); 2. Clear, obvious (to the understanding); 3. (of actions) Open, unconcealed (Glare 1982: 626). 68 These conceptual confusions are not unique to French, but can be extended to all Romance languages, e.g., Spanish or Italian, as can be seen in their respective definitions for the word evidence, which contrast with the Latin etymology of the word. 60 through diverse types of evidence. Willett (1988: 57) established a classification of these and distinguished between two major types: (i) direct evidence and (ii) indirect evidence. The former includes those sorts of evidence that come from our sensory experience, that is, our physical perception. It is characterized by being attested evidence or, in plain words, by entailing the presence of the speaker during the fact(s) considered as the source of the information. Within this type, he made finer distinctions by dividing it into three groups ranking from more to less direct evidence. These are: (a) visual, (b) auditory, and (c) other sensory, where evidence obtained via the sense of touch, smell, or taste were included. Such a distinction clearly reflects and emphasizes the heterogeneity of sense modalities and virtually coincides with Viberg’s (1984: 136) hierarchy, previously referred to in 1.2. (cfr. Figure 1.). The latter type, on the other hand, denotes that the speaker did not directly perceive what is described by the proposition. It is divided into two subcategories labeled reported and inferring. The first one, as its name suggests, encompasses those kinds of evidence that come from a person or source different from the speaker, who merely reports or quotes the information. In this group, he includes what is known as hearsay, i.e., evidence that originates from another person, the one relaying the information to the speaker, and that can be either second or third-hand, as well as evidence from tradition or common knowledge named folklore. 69 As opposed to this, in the second subcategory, inferring, the source of the information is found in the speaker’s own cognitive processes. It, then, covers two sorts of evidence, one obtained through an inference based on observable results and another acquired through deductive reasoning based on what is generally known. These are called results and reasoning, respectively. Willett’s classification is schematized in Figure 3. below: Figure 3. Willett’s (1988: 57) taxonomy of the types of evidence. 69 Palmer (1986: 40-41) finds this specific term misleading because it “usually refers to what is generally believed to be true, not necessarily as part of folklore or tradition.” TYPES OF EVIDENCE Direct Attested Visual Auditory Other sensory Indirect Reported Hearsay Second-hand Third-hand Folklore Inferring Results Reasoning 61 Willett’s taxonomy is the most influential as well as the mainstream approach in evidential typology so far. That, of course, is not tantamount to saying that it is the only one. Several scholars have offered their own view on this matter. Frawley (1992: 413) proposed another well-known classification, which was structured around a parameter different from Willett’s. Instead of basing his taxonomy on the way of acquisition of the evidence, direct vs. indirect, Frawley based it according to the source of the evidence, internal vs. external. In plain words, he constructed his classification around whether the source of the information comes from ‘me’ or from ‘other.’ Evidence is considered internal if it is obtained by means of our sensory experience or of our inferences, whereas it is considered external if it comes from hearsay or folklore. In more recent analyses, Nuyts (2001a, 2001b) and Cornillie (2007) classify evidence according to two dimensions: source-evidentiality, i.e., “the different types of knowledge involved”, and (inter)subjectivity, which refers to “the shared status of the evidence (and the proposition)” (Cornillie 2009: 45). And Squartini (2008) argues for a mixture of both Willett’s and Frawley’s taxonomies to make his own, thus allowing an interplay of the two notions. In this way, he distinguishes the types of evidence based on the mode of knowing, which includes physical perception, inferences, or reports, and the source of evidence, which can be either internal or external. Having described the types of evidence as well as some of the different taxonomies, the question of how they are linguistically expressed has to be dealt with. Every language uses different means of encoding or specifying the source of the information, which can be divided into two main groups: grammatical and lexical. For evidentiality to constitute a grammatical category, it must be expressed mandatorily in the language, a condition that is only met by “about a quarter of the world’s languages” (Aikhenvald 2004: 1) and that excludes most Indo-European ones. We find, for instance, Amerindian languages like Tuyuca or Wintu, as well as Asian languages including Korean or Japanese. Languages that do incorporate evidentiality as part of their grammar and, thus, show grammatical evidentiality typically mark it through verbal affixes and feature an evidential system as an integral part of them. These systems are not equal among languages, quite the reverse, they differ in their complexity, i.e., in how many types of evidence are marked within a certain language. They can be broadly divided into two categories (Aikhenvald 2003: 3 and ff.): (i) those which only mark the existence of a source for the information or knowledge without specifying its type (e.g., the one in Turkic); and (ii) those which do specify the type of source. This latter category can be further subdivided according to both the number and types of sources encoded by a particular language. They range from the simplest systems, which 68 evidential markers in English and there are only a few studies focusing on them (e.g., Gisborne 1998; Gisborne and Holmes 2007; Usoniene 1999, 2000; or Whitt 2008, 2009, 2010, 2011). In this section, the different evidential meanings that can be encoded by these verbs will be addressed. 75 Physical perception verbs can unmistakably be used to provide the source of information. In fact, verbs from all six semantic fields and all three parameters in Table 2. (see page 46) can be found functioning as evidential markers. In order to further explore this function, there is a vital semantic and pragmatic aspect that must be first taken into account: deixis. This term, from the Greek δεῖξις, meaning ‘demonstration,’ ‘display,’ or ‘reference,’ is roughly defined as the point of reference. Since “[evidential] utterances typically include indicators pointing directly to particular sources or away from potential sources, as the speaker takes a particular point of view in describing an action” (Joseph 2003b: 308), evidentiality can be described as being deictic by nature. For this very reason, physical perception verbs serving an evidential function must include this deictic component in addition to the original meaning that the verb in question denotes, may it be physical perception or any other semantic extension that results from metaphorization or the different syntactico-semantic processes that may be involved in its characterization. That is to say, evidentiality seems to be a built-in component of physical perception verbs. With respect to the specific evidential meanings, as predictable, physical perception verbs can indicate direct evidence, that is, they can express that the evidence for the proposition was obtained directly via the senses of the speaker: (29) a. I saw Eve leaving the house. b. I heard a dog barking yesterday. c. I smelt the cake burning in the oven. Here, the speaker is not simply stating that Eve left the house, that a dog was barking the previous day, or that the cake was burning in the oven, but also that the information conveyed comes from his own perception, in this case visual, auditory, and olfactory respectively. Examples in (29) are, then, composed of two propositions: [I saw [Eve leaving the house]2]1, [I heard [a dog barking yesterday]2]1, and [I smelt [the cake burning in the oven]2]1. 76 The 75 As Whitt (2010: 255) states, certain types of evidential meaning are structure-dependent, but within a specific type of construction more than one type of evidential meaning may be expressed as well. This is what he calls “constructional polysemy.” For a more in-depth study on this issue, refer to Whitt (2009, 2010, 2011). 76 Just in passing, in all three cases the second proposition can be viewed as an instance of a small clause whose Subject, Eve, a dog, and the cake, has features which can also characterize it as the Object of the verb of 69 second proposition carries the main information of the sentence, whereas the first one carries the evidential meaning and “is used to describe the speaker’s relationship with the [second] proposition” (Whitt 2010: 256). 77 Let us compare, however, these sentences with those in (30) below: (30) a. Brian saw Eve leaving the house. b. Lily heard a dog barking yesterday. c. You smelt the cake burning in the oven. At first glance, these examples may be thought of as being identical and conveying the same meaning as the previous ones, but by having a closer look at them, it can be noticed that such a description is not exactly accurate. Notwithstanding the structural, syntactic, and even almost lexical identity between the two sets, sentences in (30) are not considered to encode evidential meaning in contrast to the first set. The key point for this difference lies in the grammatical person of the Subject. In (29), the three sentences have first-person Subjects, whereas in (30), they have thirdand second-person ones. As mentioned above, evidentiality is deictic, which means that the speaker must be the perceiver, i.e., the Subject-referent, for the sentence to carry evidential meaning. Instances like those in (30) where the speaker is not the actual perceiver of the evidence are called “secondary stance” by Biber and Finegan (1989: 97) and, even though they are not usually described as evidential, it is worth pointing out that: These sentences conceivably could be considered evidential in contexts where the S[peaker]/W[riter] employs this secondary stance as evidence for his/her assertion […]. However, the focus here is on S/Ws’ own perceptions serving as evidence for their utterances, and such instances of secondary stance serving an evidential function fall outside the purview of the current discussion. (Whitt 2011: 350) The high degree of semantic extension the physical perception verb class displays can be extrapolated to the evidential domain as well, for the evidential meanings expressed by it are equally manifold. This means that these verbs can encode both types of evidence, direct and the first proposition, saw, heard, and smelt. This fact may give rise to two alternative syntactic analyses. On the one hand, the primary predicate could be analyzed as monotransitive, i.e., showing a /SVO/ pattern. And, on the other hand, the primary predicate could be considered complex transitive, following a /SVOC/ pattern (refer to section 3.2.2.3.). This latter possibility would further involve Subject-to-Object raising, where the Subject of the small clause would be raised to the Object position of the main clause. However, these parallel analyses would not have repercussions at the semantic level as the semantic configuration would remain the same in both cases, [Exp, Stm]. 77 Sentences like I saw Eve, I heard a dog, or I smelt the cake are out of the question for they are only composed of one proposition and only carry the meaning of physical perception. They do not encode any deictic meaning and, thereby, any evidential interpretation. 70 indirect. Since direct evidence has already been covered, it is time to focus upon the different sorts of indirect evidence that can be indicated by physical perception verbs, the first one being reported. In this category, given the specific nature of it, only experiencer-based verbs from the semantic field of hearing and sight have been attested: 78 (31) a. I heard she resigned last week. b. I heard from the news that the attacks between Israel and Hamas have not ceased yet. c. I saw in the news that the attacks between Israel and Hamas have not ceased yet. d. I heard that if you throw a coin into the Trevi Fountain, your wish will come true. Sentences in (31a-c) are instances of hearsay, the only difference between them is that in the first one the source of the hearsay is left unspecified, whereas in the other two it is made explicit. Similar to what happens with the to-phrase in verbs from the STATE parameter, the source of the hearsay is usually realized by a syntactically omissible element, being overtly expressed due to communicative purposes or the speaker’s need to emphasize it, for example, in order to provide more reliability to the information. 79 Sentence in (31d), on the other hand, is an instance of folklore, where “the original source of the information is completely unknown” (Kittilä 2020: 699) and so it is not specified. The second type of indirect evidence physical perception verbs indicate is inference. As opposed to the previous category, in this one there are no a priori restrictions on the specific type of verb that can carry this meaning. In fact, this is the only evidential meaning that can be encoded by verbs under the STATE parameter, or, in other words, phenomenon-based verbs. With regard to this group of verbs, in spite of the fact that their Subject-referent is not the perceiver or, in more semantic terms, the Experiencer, but the Stimulus, they are still viewed as carrying evidential meaning. The reason behind it is that the verbs from the STATE parameter are epistemic, i.e., they convey subjective information from the Experiencer’s own perception despite the fact that this element does not occupy the Subject position, as just stated, nor is even necessarily overtly expressed in the sentence. 80 It should be noted, nonetheless, that the Experiencer has to be necessarily the speaker and not someone else for the sentence to carry 78 The use of sight verbs in this type of evidence may be due to a fact noted by Viberg (1984: 136-137): “a verb having a basic meaning belonging to a sense modality higher (to the left) in the hierarchy can get an extended meaning that covers some (or all) of the sense modalities lower in the hierarchy.” 79 This clause constituent, therefore, is not included among the complements of the verb it accompanies. 80 The disguise of the Experiencer (or the Stimulus) goes beyond the physical perception verbs realms. For example, it can appear as a possessive pronoun in the sentence Our eyes met. 71 evidential meaning akin to the examples in (29). In this way, sentences with verbs from the STATE parameter consist of just one proposition, but “there are certainly two layers of meaning” (Whitt 2010: 256): 1) the speaker has perceived the grammatical Subject, and 2) the speaker has inferred the semantic content of the proposition based upon his own perception (the specific type being indeed determined by the actual verb). Every physical perception verb in (32) does, then, carry that deictic component —despite having secondand third-person Subjects— together with the meaning of physical perception and, obviously, the inferential evidential interpretation: (32) a. You look tired. b. The house looked expensive. c. She sounded really nervous during the conference. d. The kitchen smells as if someone had just cleaned it. 81 As mentioned previously, the STATE parameter is characterized by having a Stimulus as the Subject instead of an Experiencer, as is the case with the other two parameters. However, as pointed out in section 2.3.1., it is also possible for verbs under this parameter to have the Stimulus, the Experiencer, and the speaker conflated into that very syntactic element, the Subject, as (33) shows. Subsequently, these verbs have a first-person Subject but, contrary to what happens with experiencer-based verbs, they do not carry any evidential meaning. The reason is that, in such cases, the speaker does not need to make an inference or recourse to additional sources of information as they have immediate access to it (Whitt 2011: 350). (33) a. I look tired. b. I sound nervous. Verbs in the PROCESS parameter may also carry inferential evidential meaning as can be appreciated in the sentences below: (34) a. I see that you are a scammer. b. I saw that she had not slept at all that night. c. I heard from his tone how little sympathy he had for her. d. I smelt the cork and saw that it was very fine wine. e. I feel from the way they are talking that they like each other. 81 These are instances of raising verbs. Refer to footnote 53 in section 2.2. of the present study. 72 In (34a), the speaker visually perceives the addressee’s behavior and, based on that, makes an inference about him, that he is a scammer. Likewise, in (34b), the fact that she had not slept that night is not directly witnessed, but the observable results (dark circles, baggy eyes, constant yawning, etc.), which are used by the speaker to infer that lack of sleep. Examples like the previous two lead us to the ensuing fact: even though visual perception is described as the most objective source of information, it can also be used to indicate a more subjective type such as an inference, since everyone does not necessarily infer the same from the same evidence. On the other hand, in (34c), the speaker indicates that it is the tone of the person, perceived through hearing, that serves as the basis for the inference, as sympathy itself cannot be directly auditorily perceived. In each and every of these three sentences, the particular sense used to obtain the basis of the inference coincides with the field of the physical perception verb used to encode it. Example (34d), however, is rather peculiar in this regard because, instead of such coincidence, there is a disjunction between the two. The speaker infers the quality of the wine via his sense of smell, but the verb used to encode it, see, belongs to the semantic field of sight. It can be, then, paraphrased as “I saw that it was very fine wine based on the smell of its cork.” Finally, in the last sentence, there is more than one sense involved, since the verb feel belongs to the general field in this very case. Therefore, the speaker makes such inference based on what he feels “from the way they are talking,” that is, on what he perceives both visually and auditorily at that precise moment. In sum, two separate pieces of information can be concluded from the examples presented in this section. Firstly, as noted by de Haan (2001: 206), Gisborne and Holmes (2007: 3), and Whitt (2009: 1086; 2010: 256), there is a “tendency found among evidential perception verbs: constructions involving subject-oriented [i.e., experiencer-based] perception verbs can only be evidential with a first-person subject, whereas the object-oriented [i.e., phenomenon-based] perception verbs allow evidential readings when secondand third-person grammatical subjects are present” (Whitt 2010: 256). 82 This is a direct consequence of the deictic component innate in evidentiality as well as of the epistemic nature of the verbs under the STATE parameter. And secondly, physical perception verbs can encode other types of evidential meaning beyond the field of direct sensorial perception. 82 Mind the word ‘tendency’ here. As Whitt (2010: 276, footnote 4) also explains, this is “not a hard and fast rule.” There are, in fact, a few examples that do not follow that tendency. Nonetheless, they are “far and few between, so this rule holds in virtually all cases.” 73 CHAPTER 3. Syntax 3.1. The progressive aspect As is well known, the progressive aspect (also called durative or continuous aspect) 83 “indicates a happening in progress at a given time” (Quirk et al 1985: 197). Semantically speaking, Quirk et al (1985: 198) distinguish between three components in the progressive: (a) The happening has duration. (b) The happening has limited duration. (c) The happening is not necessarily complete. These components do not need to be all present for an instance to be in the progressive, as it is proved by the fact that they are typically assigned to different types of situations: “component (a) is distinctive for single events; whereas component (b) is distinctive for states and habits. The component of incompletion (c) is distinctive chiefly in the case of certain types of dynamic verb meaning called CONCLUSIVE” (Quirk et al 1985: 198). Given that the first two components refer to the concept of temporariness, Leech and Svartvik (2002: 52) propose a reduction of the components to just two: (a) The activity is temporary (i.e., of limited duration). (b) It need not be complete. Comrie (1976: 3) defines aspect as “different ways of viewing the internal temporal constituency of a situation” and states that subjectivity rather than objectivity is often present in aspectual oppositions (1976: 133). Hence, the above components express the speaker’s point of view regarding a specific situation, i.e., whether he or she perceives the situation in question as either temporary or permanent, or as either inconclusive or culminative. 3.1.1. Stative vs. Dynamic verbs There is a verbal dichotomy that goes hand in hand with the progressive aspect, that between dynamic and stative verbs. 84 Not all verbs can take the progressive. Those that can readily admit it are known as dynamic or non-stative; those that reject it are called stative. Semantically, the main difference between them is that whereas dynamic verbs denote an event 83 Note that Huddleston and Pullum (2012: 163) use the term ‘progressive aspect’ to refer to the syntactic form of the progressive, whereas they prefer the label ‘progressive aspectuality’ to refer to the meaning the progressive form denotes. 84 Refer to Lakoff (1966), which is the acknowledged starting point of this dichotomy. 74 that involves some sort of change, stative verbs usually denote a state that involves no change whatsoever. Using Huddleston and Pullum’s (2012: 119) words: “states exist or obtain, while occurrences [i.e., events] happen, take place. Occurrences involve change, while states do not. States have no internal temporal structure: they are the same throughout their duration, having no distinguishable phases.” This way, states are contrary to the semantic components inherent to the progressive aspect. Focusing on stative verbs, they have been subdivided into different groups usually based on semantic grounds. Quirk et al (1985: 200-201) draw a distinction between (i) verbs denoting qualities, which are conceived as “relatively permanent and inalienable properties of the subject referent” (e.g., Mary is Canadian); and (ii) verbs denoting states, which “introduce the less permanent situation types” (e.g., Mary is tired). To these two categories, they (1985: 205-206) also add a third one called “stance,” which represents a small class of verbs and stands in an intermediate position between dynamic and stative verbs. It appears in such a middle position because the members of this category (e.g., live, stand, lie, or sit) can occur in the progressive implying a temporary state (35a) or in the non-progressive implying a permanent state (35b): (35) a. James is living in Copenhagen. [temporary residence] (Quirk et al 1985: 206) b. James lives in Copenhagen. [permanent residence] (Quirk et al 1985: 206) Among the large array of verb classes, there are some that have been traditionally described as stative. Quirk et al (1985: 203) list the following: (a) Intellectual states (e.g., believe, know, think). (b) States of emotion or attitude (e.g., intend, like, want, wish). (c) States of perception (e.g., feel, hear, see, smell, taste). (d) States of bodily sensation (e.g., ache, hurt, itch, tickle). 85 This taxonomy is shared by many linguists, such as Leech and Svartvik (2002: 53), who also add one more category to the prior four ones: that of verbs referring to a relationship or state of being (e.g., be, belong to, concern, consist of, owe, own). However, this classification is not without debate, as there are some linguists who do not see the distinction between stative and dynamic verbs so clear-cut. To take one example, Huddleston and Pullum (2012: 168-169) 85 This last verb class, i.e., verbs of bodily sensation, can be used more or less interchangeably with the progressive or non-progressive when referring to a temporary state. Then, these two sentences (Quirk et al 1985: 203) are correct, since they imply a temporary pain, and not a permanent state: (ii) a. My foot hurts. b. My foot is hurting. 75 claim that it is not possible to draw a sharp line between states and occurrences in objective reality and that, consequently, some verb classes fall around the boundary. The specific verb classes they (2012: 169-171) refer to are: (a) Verbs of perception and sensation (e.g., see, hear, feel, smell, taste). (b) Verbs of hurting (e.g., ache, hurt, itch). (c) Verbs of cognition, emotion, and attitude (e.g., believe, fear, regret, intend, like). (d) Stance verbs (e.g., stand, lie, sit). Interestingly enough, verb classes from (a) to (c) coincide with those classified as stative by Quirk et al and, except for one, with the classification by Leech and Svartvik. This boundary position is, then, composed by all four Quirkian verb classes, (a) and (b) grouped together into just one class, (c), and stance verbs, a fact already noted by Quirk et al (1985: 205-206). Accordingly, there are several instances in which the members of these verb classes appear in both the progressive and the non-progressive without being grammatically incorrect: (36) a. I love you. b. I’m loving it. (37) a. I think it’s impossible. b. Don’t interrupt me when I’m thinking. (Huddleston & Pullum 2012: 170) Such a lack of a consensus about which verbs are stative and which verbs are not shows how difficult it is to generalize about verb classes belonging to a particular category or behaving in a particular way. The inclusion of a particular verb class in the stative category depends on the meaning of the different verbs that compose it, but verb meaning and behavior can be as varied and heterogeneous within an individual verb class and even verbal form as it can be among the diverse verb classes (refer to section 2.1. of the present study). A verbal unit may seem to denote a state in a particular context, whereas in another it may seem to denote an event. This means that, although linguists such as Quirk et al (1985: 203) and Leech and Svartvik (2002: 53) classify these verbal forms as stative, their alleged incompatibility with the progressive is not completely clear-cut. Needless to say, these linguists are not unaware of the aforementioned examples and, therefore, have tried to provide an accurate explanation for them, so that their classifications are still valid. Quirk et al (1985: 202) point out that when stative verbs are combined with the progressive, a change of interpretation is needed, usually involving a reclassification of the verb as dynamic, “e.g., as having a meaning of process or agentivity.” Leech and Svartvik (2002: 76 53) also support this view of reclassification and explain that in those situations the verb refers to an “active form of behaviour.” All in all, the progressive construction should not be regarded as just a phenomenon purely grammaticized in structure. It can —and actually does— house pragmatic features that are transferred to the whole of the proposition, such as the two components aforementioned: temporariness and incompleteness. Therefore, the progressive cannot be used with any verbal unit at random, whether a verb admits or rejects the progressive has to do with its specific meaning in a specific context. 3.1.2. Physical perception verbs and the progressive: Their interaction As has already been mentioned, physical perception verbs are traditionally included in the stative category and, thus, their co-occurrence with the progressive is deemed incompatible or at least unorthodox in most grammars. Several arguments have been put forward to justify such a behavior, which is considered primary or central of stative verbs. Leech and Svartvik (2002: 53) claim that they do not as a norm occur in the progressive, “because the notion of ‘something in progress’ cannot be easily applied to them.” Quirk et al (1985: 198) make a stronger assertion by expressing that stative verb meanings are “inimical to the idea that some phenomenon is ‘in progress’.” Along the same lines, Comrie (1976: 35) states that: [V]erbs tend to divide into two disjoint (nonoverlapping) classes, those that can appear in the progressive forms, and those that cannot. Moreover, this distinction corresponds to that between stative and nonstative verbs. Thus we can give the general definition of progressiveness as the combination of progressive meaning and nonstative meaning. Naturally, then, stative verbs do not have progressive forms, since this would involve an internal contradiction between the stativity of the verb and the nonstativity essential to the progressive. Nevertheless, as pointed out by authors such as Huddleston and Pullum (2012: 169-170) or Jørgensen (1990), there are numerous instances in which physical perception verbs occur in the progressive construction without that affecting their grammatical correctness, a clear case being those with verbs belonging to the ACTION/PROCESS parameter (see Table 2.). Many authors have addressed this issue and offered an explanation for such ‘rare’ or ‘exceptional’ cases. Hornby (1975: 103) claims that perception verbs cannot occur in the progressive because states or feelings are assumed to have no end and cannot be started and stopped at will. Thus, when perception verbs are used in the progressive, contrary to the rule, what happens is that they undergo a change of meaning or some sort of specialized meaning that sets the core 77 meaning apart. That is to say, these verbs acquire a different semantic reading which is no longer stative, but dynamic, allowing them to occur in such a construction: (38) a. I’m seeing my dentist this afternoon (= paying a visit). (Hornby 1975: 103) b. Which judge is hearing the case? (= try). (Hornby 1975: 104) The change of meaning may not only involve a change of category from stative to dynamic, but also a change of parameter from PROCESS to ACTION/PROCESS (see Table 2.) in order to account for the agentive character of the new meaning. This view is akin to that by Quirk et al (1985: 205), who, focusing on this verb class, propose that “the exceptional occurrence of the progressive [with physical perception verbs] seems to arise from perception being treated as an ongoing process.” In other words, physical perception verbs can occur in the progressive construction when they show a feature typical of the progressive aspect and, thus, can be reclassified as dynamic: (39) I need some spectacles. I’m not seeing things so well these days. (Quirk et al 1985: 205) Likewise, Palmer (1988: 71) suggests that the progressive is only used with this verb class when “there is specific reference to duration or one of the special features indicated by the progressive.” In addition to this, he (1988: 71-72, 74-75) avers that there are other explanations for those “more unusual” instances to which the former reason does not apply. Physical perception verbs can occur in the progressive construction when the meaning indicates a change of state, as in (40a), or repetitions of the sensation, as in (40b), which is normally imaginary or hallucinatory: (40) a. It’s tasting nastier and nastier. (Palmer 1988: 72) b. She’s hearing voices. (Palmer 1988: 75) Leech (2004: 25) goes a bit further and states that these cases in which a stative verb can appear in the progressive can be explained “by supposing that the verb (perhaps through a special transfer of meaning) can become a member of a different verbal category.” He (2004: 28) illustrates this multiple verb class membership by means of the verbal form feel, which, as he claims, can stand for verbs that belong to four different verb classes depending on the meaning: firstly, to ‘verbs of inert perception’ when expressing external sensation (41a); secondly, to ‘verbs of inert cognition’ when denoting the speaker’s feeling or opinion (41b); 84 3.2.1. Intransitive Intransitive verbs subcategorize one single element, the Subject, as their only completing structure, thus showing a /SV/ pattern. Nevertheless, a word of caution should be made here, as we should distinguish between pure intransitive verbs and pseudo-intransitive verbs. The former type represents the standard conception as described above. The latter type, on the contrary, stands for those verbs which are normally or prototypically monotransitive, but whose Object, though semantically present in the core meaning of the verb, is not syntactically explicit due to its lack of communicative relevance. That is to say, there is no univocal relationship between the syntactic and the semantic levels of analysis. In the physical perception verb class, both types have been recorded, albeit not in every parameter. Although pure intransitive verbs have been found in all three parameters, pseudointransitive verbs have been found in only two of them, namely the PROCESS and ACTION/PROCESS parameters. This may well be motivated by two related reasons: (i) the semantic content of these verbs, i.e., their meaning, and (ii) the fact that, as a consequence of (i), the verbs under the STATE parameter have not been attested functioning transitively, blocking, this way, their behavior as pseudo-intransitive verbs. When the verbs belonging to the PROCESS parameter behave as pure intransitive verbs, the most typical meaning is to have the capacity or the ability to use that specific sense —or senses when dealing with verbs from the general field. Hence, it entails that the Subject-referent unwillingly and involuntarily, i.e., non-agentively, has such a capacity of perceiving through one or more of their senses: (50) a. He can’t see after the accident. b. Why can’t we smell or taste when we have a cold? In this parameter, there are also extended and metaphorical meanings which correspond to pseudo-intransitive verbs, such as ‘to understand’ (51a), ‘to find out’ (51b), or ‘to be adjacent to, to border on’ (51c). These three verbs are also found functioning as monotransitive verbs, as will be seen, and, in fact, they are related to them through the deletion of their Objects, which are contextually recoverable and understood by the participants of the communicative exchange. In the case of (51c), touch is also functioning as a symmetrical verb as in examples in (24) (see section 2.3.2.): 85 (51) a. –She didn’t want to ruin the surprise, so she didn’t say anything. –I see. b. –Can you climb up that tree? –We’ll see. c. Push the two dressers until they touch. Likewise, the verbs under the ACTION/PROCESS parameter may behave intransitively and pseudo-intransitively. The latter behavior has been recorded when they convey the most prototypical meaning, to willingly and voluntarily perceive something through one or more of the senses: (52) a. Are you going to help me with this or just watch? b. Let’s be quiet and listen first. As mentioned above, in the STATE parameter the situation is the opposite. That is, only pure intransitive verbs have been recorded. This pattern appears in five of the six semantic fields of physical perception verbs: sight, hearing, smell, taste, and general. In addition to this, there are two peculiarities: The first one is that the Subject can be realized by an as if/though clause or a like clause, but, interestingly, they do not usually appear in subjective position due to the process of extraposition, which is compulsory in this case. 91 Through this postponement or rhematization movement, an expletive or anticipatory it is inserted in the Subject slot and the real or logical Subject of the sentence is dislocated or extraposed to the typical post-verbal position of other completing structures, say, the Object or the Complement. Thus, its apparent structure is that of an adjunctive verb, /SVA/. The role of the anticipatory it is “essentially a structural one in the sense that it carries virtually no information in itself, but merely supplies the structural requirement for an initial subject” (Quirk et al 1985: 89). (53) a. It looks as though it’s going to rain. b. It sounds as if they had a good holiday. c. It felt as though he had run a marathon. d. It looks like he is going to win. e. It sounds like I’m criticizing you. f. It feels like I don’t belong here. 91 In many other cases, extraposition is optionally applied and there is another alternative structure. Which one to use is motivated by the end-weight principle and/or the end-focus principle. The former stipulates the tendency to place longer, more complex, or “heavier” structures at the end of the sentence, and the latter establishes the tendency to place the focus of the information, i.e., the most informative element, at the end of the sentence. 86 The second peculiarity is that when the verb functioning intransitively is smell, that verb may acquire negative connotations similar to those of the verb stink, which belongs to the same semantic field and parameter, and can also behave intransitively. The only difference between them is that the latter implies a stronger negative or unpleasant sense. When it is used in a negative construction, this connotation disappears, taking a neutral one: (54) a. Your feet smell. = Your feet stink. b. This flower does not smell. ≠ This flower does not stink. 3.2.2. Transitive 3.2.2.1. Monotransitive This syntactic behavior reflects the prototypical pattern in the PROCESS and ACTION/PROCESS parameters, namely /SVO/. What is more, it coincides with the traditional view which states that physical perception verbs just require a Subject and an Object, those syntactic elements that correspond to the two participants involved in the situation described by the verb, somebody to perceive and something to be perceived. This way, this complementation pattern is attributed to the most prototypical meaning of physical perception in both parameters, i.e., to physically perceive something with one or more of the senses, either involuntarily (55) or voluntarily (56): (55) a. Emma smelt sulfur. b. The neighbors heard a loud noise last night. c. I noticed a spark in her eyes that day. (56) a. I tasted the tea to see if it was sweet enough. b. Have you watched Star Wars: Ahsoka? c. Can you feel whether there are any bones broken? Apart from this, there are many other extended and/or metaphorical meanings which also behave in this fashion. Regarding the PROCESS parameter, we find the monotransitive counterparts of those verbs in (51), namely ‘to understand’ (57a), ‘to find out’ (57b), and ‘to be adjacent to, to border on’ (57c), as well as other meanings such as ‘to suspect’ (57d),‘to be affected emotionally’ (57e), or ‘to be told, to be informed’ (57f): (57) a. He finally saw what she meant. b. We’ll see if you can finish this project on time. c. Their hands were touching each other. 87 d. She might have smelt that there was something wrong about it. e. My father felt having to sell our old house. f. –She went to Edinburgh last month. –So I heard. In the ACTION/PROCESS parameter, we also find, for instance, ‘to pay attention’ (58a), ‘to examine’ (58b), ‘to show contempt’ (58c), or ‘to plan’ (58d). (58) a. We’d better hear what they have to say. b. I’ll see whether the proposal is viable or not. c. They asked her, but she sniffed that it was the only way to do it. d. We are looking to sign a contract with a big company. As can be seen in examples (55)-(58) above, the Subject is usually realized by a NP, as it normally refers to a person, whereas the Object shows more flexibility and may be realized by several different elements ranging from simpler to more complex structures. The simplest and usual possibility is a NP, as in (55) and (56a-b), but it is also fairly common to find instances in which the realization of this particular syntactic element is a more complex structure, namely a clause. The nature of the clause itself varies depending on the communicative intent on the speaker’s part and on the parameter. That is to say, although this is a shared feature between the verbs from both parameters, the types of clauses that these can subcategorize differ. The physical perception verbs within the PROCESS parameter can select as their Object a whclause (57a), an if/whether clause (57b), a that-clause (57d), an –ing clause (57e), 92 or a so proclause (57f). 93 The ones from the ACTION/PROCESS parameter can also select a whclause (58a), an if/whether clause (58b), and a that-clause (58c), but they cannot select an -ing clause or a so proclause. Instead, their Object can also be developed by a to-infinitive clause (58d). 3.2.2.2. Ditransitive The simple ditransitive pattern, i.e., /SVOO/, is less usual within this verb class and is restricted to a specific meaning in the hearing field of the ACTION/PROCESS parameter. This pattern is only found when the verb hear means ‘to hear sb recite sth (usually his/her lessons)’ and both Objects, the Indirect and Direct one, are developed by NPs. It is worth mentioning that this verb does not allow the permutation of neither of its Objects through the prepositional construction 92 This clause-type has also been referred to by some linguists as “gerund clause” (Kreidler 2014: 99) or “gerund-participial clauses” (Huddleston & Pullum 2012: 1220). 93 So is a proform of a that-clause. 88 /SVOPO/, a syntactic alternative which is normally admitted by ditransitive verbs. This specific permutation process is called “dative alternation” by Levin (1993: 45): (59) a. Today I cannot hear Mary her lessons. (Garrudo Carabias 1996: 240) b. *Today I cannot hear her lessons to Mary. 3.2.2.3. Complex Transitive The verbs sheltered under the PROCESS and ACTION/PROCESS parameters can also require an Object and a Complement, showing an /SVOC/ pattern and becoming complex transitive verbs. These two elements show an intensive relation in which the Complement predicates something about the Object, i.e., they are co-referential, and, therefore, the Complement is, more specifically, an Object Complement (CO). Regarding their formal realizations, the Object is typically performed by a NP in both parameters. In some cases, it is possible to find a reflexive pronoun in that position, as in (60) below, creating a reflexive construction. That is to say, the Object acts as a reflection (hence, ‘reflexive’) of the Subject-referent: (60) a. My cousin sees himself winning the race. b. You can watch yourself perform on TV if you pass the audition. 94 However, whereas there are no parametric differences in the nature of the first element, the nature of the second one does vary. In the PROCESS parameter, the Complement can be realized by either an AdjP (61a), a NP (61b), or a clause of different typologies, an –ed clause (61c), an –ing clause (61d), or a bare infinitive clause (61e): (61) a. We love to see you so excited about your new job. b. She felt it her duty to tell the police. c. They saw a man chased by policemen. d. I could smell the food burning in the kitchen. e. She felt the razor touch her throat. (Garrudo Carabias 1996: 160) In the case of the ACTION/PROCESS parameter, it can only be developed by a clause, specifically, an -ed clause (62a), an –ing clause (62b), or a bare infinitive clause (62c): 94 These two examples could be conceived as including small clauses, where the reflexive pronoun, himself/yourself, would be the Subject of a secondary predicate, winning/perform: My cousing sees [himself winning the race]SMALL CL/ You can watch [yourself perform on TV]SMALL CL. This would involve the analysis of the small clause as the Object of the primary predicate, see/watch, and, as a consequence, the analysis of these verbs as monotransitive. Refer to footnote 76 as well as to the information in this very section. 89 (62) a. We only looked at the car destroyed after the accident. b. He smelt the cake baking in the oven. c. The whole class listened to him recite his poem. There are two aspects regarding this model of syntactic complementation which need to be discussed. The first one is related to one of the syntactic elements, the Complement, and has repercussions on the interpretation of the sentence; and the second one affects the syntactic description of these verbs and structures, resulting in alternative analyses. Among the types of clauses that can function as Complement, there are two which have received special attention and have been studied contrastively. These constructions correspond to two non-finite clauses, the –ing clause and the bare infinitive clause. The reason lies in the fact that, when they both can be admitted by the same verb in the same linguistic context, the use of one or the other leads to distinct interpretations. To put it differently, physical perception verbs can admit both clause types as their Complements, yet the choice of one of them is not arbitrary by any means; on the contrary, they imply different semantic and pragmatic readings at the level of aspectual relationships. The –ing clause expresses an action in progress that is not perceived in its totality, whereas the bare infinitive clause indicates that the action is perceived to its completion. In other words, “the perceived event begins either before, continues and ends (or not) simultaneously with the moment of perception, or begins and ends with the act of perception” (Manasia 2014: 330). To illustrate this, let us consider the following pairs of examples: (63) a. I saw him writing the letter. b. I saw him write the letter. (64) a. A teacher observed her climbing over the gate. b. A teacher observed her climb over the gate. Despite each pair of examples in (63) and (64) being constituted by exactly the same elements in the same order, they cannot be used interchangeably, for they do not share the same semantic or communicative meaning. The ones with an –ing clause functioning as the Complement convey the meaning that only part of the process of writing or climbing was perceived, without specifying whether those processes were perceived until they were completed. Then, the emphasis is on the progress or development rather than on the completion of the process. Contrarily, the ones with a bare infinitive clause realizing the Complement imply that what was perceived was the process of writing and climbing in its totality, that is, from the beginning until the end. This way, the process is viewed as a whole, involving a beginning, a 90 development, and an end, with an emphasis on its completion. In light of all this, it could be stated that whereas the first one has a partitive effect, the second one has a holistic effect. This question has also been studied in terms of another familiar opposition which has already appeared in this study: stativity vs. dynamicity. Van der Meer (1994: 477) explains that the fundamental difference between both types of structures when functioning as Complements lies on these two notions. He (1994: 477) states that “the construction with ing gives a STATIC [sic], frozen, picture whereas that with bare infinitive gives a DYNAMIC [sic] picture”. Such a connection is particularly interesting as the –ing structure is described as static and the bare infinitive as dynamic, which contrasts with the fact that when a verb allows the –ing structure is called dynamic and when it rejects it, stative. The second aspect that needs to be discussed before going any further is that there is some controversy about whether these specific constructions can be analyzed as following a /SVOC/ or a /SVO/ pattern. Although it is generally assumed that there is an Object and a Complement, according to several linguists (Quirk et al (1985), Van der Meer (1994), Felser (1999), and Radford (2004), inter alia), these two elements can be envisaged as being a unified and single constituent, namely an Object, functioning as a small clause (Gómez Castejón 2007). Opinions also differ depending on the specific clause being discussed. Whereas there seems to be a greater support for the /SVOC/ analysis, i.e., the description of these verbs as complex transitive —rather than monotransitive—, when dealing with an –ing clause, the same is not so clear when the construction in question has a bare infinitive clause. One of the reasons for this is that the process of passivization is used as a test to prove whether these can be considered a single constituent or two separate elements. Following this, it is claimed that the construction with the –ing clause can undergo such a process, but that the construction with the bare infinitive cannot: (65) a. People saw her leaving the town. b. She was seen leaving the town. (66) a. People saw her leave the town. b. *She was seen leave the town. The underlying rationale for that forbiddance is that these elements that follow the verb in (66a), her leave the town, cannot be separated, since they make up just one constituent, the Object. However, there is a relevant factor which may have been overlooked in that reasoning and which may make this argument invalid, as it can demonstrate that, actually, both constructions allow their appearance in the passive voice. As pointed out by Quirk et al (1985: 91 1205) and Van der Meer (1994: 474), among others, when a construction with a bare infinite clause functioning as a Complement is passivized, the bare infinitive turns into a to-infinitive. 95 Thus, the passive counterpart of (66a) above is not (66b), but (67) below: (67) She was seen to leave the town. This evidence is used by those linguists, such as Quirk et al (1985), Langacker (1991), and Van der Meer (1994), who hold the view that both constructions are composed by two separate constituents and follow the same /SVOC/ syntactic pattern. Furthermore, Van der Meer (1994: 475) explains that both constructions, that with a bare infinitive clause and that with an –ing clause, can appear in a coordinated structure, implying that “there is at least a sufficient degree of structural similarity between the [bare] infinitive structure and the -ing structure.” (68) …bearded celebrity, sometimes even having the privilege of seeing him play [sic] on the close and smiting [sic] the ball for six. A heavenly spectacle! (Van der Meer 1994: 475) Apart from these syntactic reasons, Goméz Castejón (2007: 36) expresses that there is also semantic and cognitive evidence to consider these elements as two separate constituents. From a semantic point of view, People saw her leave/leaving the town entails that People saw her. Thus, the Object can be viewed as that single element, the NP. Regarding the cognitive evidence, she uses Langacker’s (1991: 442-443) analysis and explains that the object of perception, i.e., what is perceived, is not merely a thing, or person in this case, her, “but includes some process in which that particular thing [or person] participates” (Gómez Castejón 2007: 36), leave/leaving the town. This dual analysis also has consequences or implications at other linguistic levels. That is to say, regarding these constructions as being one or two distinct constituents does not only affect their description at the syntactic level, but also at other levels, e.g., morphological. In those terms, the -ing form could be conceived as a gerund or a participle. Using Gómez Castejón’s (2007: 36) words: 95 Not all perception verbs that take a bare infinitive as their Complement allow this passivization, e.g., watch or listen (He watched the woman come out → *The woman was watched to come out; We listened to him recite his poem → *He was listened to recite his poem). Although Quirk et al (1985: 1205) also include feel, it has been recorded in a passive construction behaving as a complex transitive verb: (iii) I felt the ground shake. → The ground was felt to shake. 92 The treatment of this construction (NP and an –ing form) as two different constituents implies that the –ing form functions as a participle, where the NP is the direct object of the main verb and the –ing form is an adjunct depictive of the NP; whereas the interpretation of this construction as a single unit would involve an – ing functioning as a gerund, the construction “NP+-ing” as a unit functions as a direct object of the main verb. 3.2.2.4. Adjunctive Transitive Adjunctive transitive verbs require two subcategorized elements apart from the Subject, namely an Object and an Adverbial, showing a /SVOA/ pattern. There are few physical perception verbs that function as adjunctive transitive verbs. As a matter of fact, this model of complementation has only been attested in the field of sight, more specifically, in the verbal forms see and view, in the PROCESS and ACTION/PROCESS parameters. The particular meanings that show this behavior are ‘to regard, to consider’ (69a-b), ‘to accompany, to escort’ (69c), and ‘to meet’ (69d): (69) a. I see things in a different light now. b. How do you view the future? (Garrudo Carabias 1996: 767) c. Could you, please, see him to the door? d. I saw your wife around the corner. (Garrudo Carabias 1996: 583) It is important not to confuse an adjunctive transitive verb having a PP as the Adverbial with another type of verb, more specifically, with a prepositional ditransitive one. At first sight, they may seem similar, for they both are followed by an Object and a PP, but the relation between the verb and the preposition in the two complementation patterns is completely different. In the adjunctive transitive model, the preposition is not determined by the verbal unit, i.e., it is not, say, fixed, or subcategorized by the verbal unit. Contrariwise, it is in paradigmatic relationship with other prepositions. Therefore, the same verbal form and even the same verb can be found taking PP complements introduced by other prepositions, as can be seen in (70): (70) a. I saw your wife around the corner. b. I saw your wife at the coffee shop. In the prepositional ditransitive one, however, the preposition that introduces the PP is ultimately determined by the verb. Furthermore, both the verbal unit and the preposition are considered a whole in the sense that the preposition is fixed and cannot be replaced by another one, as in (71). If that happened, either the sentence would be grammatically incorrect as in 93 (71a-b), or the verb would convey a completely different meaning and, thus, would be regarded as a different verb, as in (71c-d): (71) a. He saw gentleness in what she did. b. *He saw gentleness into what she did. c. He saw Susan about her house. d. He saw Susan at her house. 3.2.3. Copulative Another syntactic pattern that these verbs can show is /SVC/, which corresponds to copulative verbs. The Complement, typically realized by an AdjP or a NP, refers back to the Subjectreferent by ascribing a quality to it. This means that the two arguments required by the predicate are necessarily co-referential and show an intensive relationship. Hence, this Complement is further described as a Subject Complement (CS). As pointed out in Table 2., this is the prototypical pattern of the verbs sheltered under the STATE parameter. Nevertheless, this does not mean that it cannot be found in any of the other two. In fact, a few instances pertaining to the general field in the PROCESS parameter have been recorded: (72) a. You’ll feel better after a good night’s sleep. b. Standing there on stage, I felt a complete idiot. With respect to the STATE parameter, the Complement is realized by the typical elements as well, an AdjP or a NP, yet the semantic fields of smell and taste do not, as a rule, admit a NP realizing that argument: (73) a. He doesn’t look his age. b. Her new song sounds a bop! c. The cake smells delicious. d. Caffeine is one of the reasons why coffee tastes bitter. e. Your hands feel cold. f. It feels good to be rich. g. These gloves felt a good choice, but they ended up breaking quite easily. Example (73f) represents an interesting case. The real Subject is not the one in the Subject position, but the to-infinitive clause to be rich after the Complement. That is, the original sentence is to be rich feels good. This means that we are dealing with a case of extraposition 100 rhetorical, metatextual stance towards the development of the discourse and its cohesiveness (Cuenca & Marín 2000: 234; Traugott & Dasher 2001: 155). And on the other hand, it is also due to an increase in intersubjectivity, since these DMs are also used interpersonally, for instance, as markers of negative politeness to save the addressee’s face by virtue of their less face-threatening potential. These are the so-called mitigators or hedges in Politeness Theory (Brown & Levinson 1987: 145). b) They have a parenthetical character and are positionally flexible within the sentence structure in some contexts. It would be hasty, however, to assume that such flexibility is a shared feature by each and every DM, as there may well be —and, in fact, there are— exceptions to the rule. Echoing Martín Zorraquino (2008: 44), “cuanto más individualizado se hace el studio de los marcadores, la supuesta ‘versatilidad’ o ‘movilidad’ distribucional de éstos queda más en entredicho.” 100 They appear between pauses as an independent group and normally occupy a marginal or peripheral position in the sentence, either in the left periphery (sentence-initial) or in the right periphery (sentence-final), or a medial position. The specific position is influenced by (i) the actual function the DM serves in a specific context and (ii) the construction it derives from (Mithun 2020: 27). All in all, these forms appear syntactically, prosodically, and graphically detached from the rest of the sentence, only being linearly adjacent to it in a way analogous to disjuncts and functioning at an extra-sentential level. c) They commonly co-occur with vocatives as well as in combination with other connectors or DMs which reinforce their discourse meaning and function. These elements can be of any type, e.g., adverbs such as now or here, or can arise from other physical perception verbs too. With respect to this latter option, we have attested numerous combinations of DMs evolving out of different physical perception verbs. Three of them are particularly interesting, namely ‘look, listen,’ ‘look, see,’ and ‘see, listen.’ These combinations, which show no fixed order between the two DMs, readily evince that expressed in a), the loss of lexical meaning in favor of a pragmatic one. As a matter of fact, it is that loss that enables the co-occurrence of two DMs which evolve from different semantic fields of physical perception, specifically those of sight and hearing: 101 100 The more individualized the study of markers is, the more questionable their assumed distributional ‘versatility’ and ‘mobility’ becomes (my own translation). 101 It is worth mentioning that all three combinations were registered in American English, but not in British or Canadian English according to three main corpora: COCA, BNC, and Strathy. Furthermore, the search yielded a very low rate frequency in every combination: 0.02, 0.07, and 0.01 instances per million words, respectively. 101 (79) a. Look, listen, I’ll be direct, but I’m not here to make new friends. b. Listen, look, I don’t really want to go down the regrets and recriminations. c. Look, see, Jen, it was meant to be. d. See, look, baby, this is what I’m talking about. e. See, listen, you have to let kids expand a little bit. f. Listen, see, I’ll send it to you. 102 d) There is freezing of form. As a consequence of the process of grammaticalization, these forms lose the morphological variability innate in the verbal paradigm (mind, of course, defective verbs) and, therefore, as DMs they remain frozen in a particular form or formal structure whose elements are also fixed (though other DMs can be added as noted above). These frozen forms include: • Imperative form: hear, look, listen, see. • Imperative form + second-/third-person pronoun: look you, lookit. • First-/second-person pronoun + present tense form: I/you see, I hear, do you hear?, I feel. • Adverbial finite clause: as/so you see, as I see it. • Periphrastic construction with future value: let’s see. e) Apart from losing their morphological variability, they also experience the loss of their argument-taking properties. Provided they no longer have a predicative or syntactic function but a discourse one, they lose their capacity of complementation and do not select their normally obligatory arguments, i.e., they are syntactically defective. That being said, there are instances in which these DMs do appear in combination with a Subject or Object pronoun, as mentioned in d) above, resembling more prototypical constructions. Nonetheless, this argument-taking property is nothing but apparent, as we must not forget that these are fixed elements which cannot be freely substituted. f) Finally, there are instances of morphological bonding and phonological attrition, which represent extreme cases of grammaticalization. These processes have been attested in lookforms, namely in look you, look you here, and look to it, and have not taken place simultaneously, but rather gradually on various stages. Firstly, there is phonological attrition of the second element of each construction, [yu] or [yeɪ] > [yə], and [tuɪt] > [tuət] > [tut], with its corresponding graphic representation, look ye, look ye here, and look to’t. Besides, 102 Examples in (79) are taken from COCA. 102 morphological bonding also takes place within the prepositional structure of the latter case. Secondly, there is further phonological attrition of the second element, [lʊkji] > [lʊki] and [lʊktɪt] > [lʊkət/lʊkit]. 103 And lastly, there is morphological bonding between the look-form and the other elements of the construction, which graphically reflects the prior phonological attrition and results in the forms look’ee/looky, lookyhere/lookahere, and lookit (Brinton 2001: 193; 2008: 187; Brinton & Traugott 2005: 138). 104 In terms of their functions, these DMs display a wide array of discourse functions, the particular type(s) being determined by the situational and communicative context in conjunction with the nature of the forms themselves. As mentioned above, these range from calling the addressee’s attention, encoding politeness, or managing turn-taking to others more oriented to the thematic progression of discourse like elaborating an argument or counterargument, or initiating or changing the topic. Let us develop this point further and discuss the different functions each DM under study can serve as well as their distribution within the proposition. To do so, three main dialects will be examined, namely British, American, and Canadian English, by means of three well-known reference corpora, the British National Corpus (BNC), the Corpus of Contemporary American English (COCA), and the Corpus of Canadian English (Strathy), respectively. 105 3.3.1. See-forms There are several DMs that evolve from the verb form see. You see is the most frequently used DM within all see-forms. It is claimed to be typical of British English and to have a lower frequency in American English, there being a tendency to use other DMs such as you know in those contexts in which British English would use you see instead. 106 In fact, Biber et al (2021: 1092) and Algeo (2006: 213) have reported that you see is eight times more frequent in British than in American English, representing c. 400 occurrences per million words in British against c. 50 occurrences per million words in American English. Data from the BNC and the COCA 103 The phonetic transcriptions in f) are the ones originally used by Brinton (2001: 193). 104 Mind that there is another form lookit, derived from look at, which functions as a predicative unit with a physical perception meaning. This form can be easily differentiated from the DM lookit, because it is followed by the Object. (v) Lookit that dog. 105 We did not use the latest version of the BNC corpus, BNC2014, as it does not include punctuation. This would make the search for parenthetical forms a very difficult task provided the high frequency of this verb class. 106 Just in passing, you know is twice more frequent in American than in British English (Biber et al 2021: 1091; Algeo 2006: 213). This difference may well be partly due to the preference for this DM over you see in the same contexts. 103 also support this difference or preference, as it shows a frequency of 18.27 instances per million words in British English in contrast with 11.42 instances per million words in American English. You see can serve a variety of functions. The most general one is that it can be used to claim the addressee’s attention, as already noted by Quirk et al (1985: 1115) and as it is customary in this group of DMs. Echoing Fitzmaurice’s (2004: 441) words, this is “a sure sign of an interactive discourse marker.” It encodes instructions on the information structure of the utterance, signaling that the constituent(s) on which it operates are new information and contribute to the progression of discourse (Blakemore 1987: 89). It does not assert, however, the truth-value of the proposition, i.e., it is a neutral marker (Brinton 2008: 160). Another general function is that it indicates transitions between information in discourse (Algeo 2006: 213). It is found in argumentative contexts, in which it marks an explanation for or justification of a previous claim. In this way, it works in a similar way to a connective (‘the reason is’) (Aijmer 2004: 263). That function corresponds to the usual description appearing in dictionaries. For instance, the OED (s.v. you see in see (v.), sense P.3.a.i) expresses that “[i]n parenthetical or introductory phrases with you (or †thou) as subject, used to reinforce an explanation, argument, etc., in expectation of assent or understanding.” The last part of the previous definition is especially relevant as it brings us to the next function, claiming mutual understanding. This function is related to the cognitive sense present in the verb see, ‘to understand.’ By using this DM, speakers seek the addressees’ agreement and try to persuade them to accept their arguments through negotiation, thus attaining or maintaining intersubjective alignment. In this regard, it also functions as a means of softening epistemic stance and marking politeness, since it avoids a potential face threat whenever the speaker’s utterance “might provoke surprise or blame (i.e., be a face-threatening act)” (Brinton 2008: 159) and assures that the addressee is not being underestimated by establishing common ground (Quirk et al 1985: 1481-1482). Hence, it could be replaced by something along the lines of ‘I’m sure you understand that’ (Fitzmaurice 2004: 439). It can also be found having a triumphant sense meaning ‘I told you’ (Quirk et al 1985: 1483, note [b]; Brinton 2008: 160). Furthermore, it is possible to find this DM followed by that very expression. Finally, you see may also serve as a way of eliciting agreement or a backchannel response from the addressee and to manage turn-taking, especially when occurring in an interrogative structure (you see?), with rising intonation, and/or sentence final. This fact supports the tendency —not the rule— that the right periphery is likely to correspond to interpersonal or intersubjective functions (Beeching et al 2009; Modicom & Duplâtre 2020: 9). 104 (80) a. My grandmother, you see, was from the South. b. You see, I didn’t want to bother you again. c. Everyone has a secret, you see. d. He did come, you see. I told you. e. And then he left without saying anything, you see? The DM see is not as commonly used as you see. In fact, when comparing their frequency, it can be clearly noticed that the figures are lower in general. A notable exception is attested in American English, in which this DM represents 10.9 occurrences per million words when appearing sentence initial. This means that in American English, see in initial position even outstrips you see, which is the most frequent DM within see-forms overall, regardless of the distribution or dialect. See derives from the imperative form, which originally has an inherent conative and phatic value, that is why it clearly serves as an attention-getting device. It bears a striking resemblance to you see in terms of their functioning, which may well be the reason why the OED considers the DM see a reduced form of it. This way, see is also used to provide an explanation or justification for a preceding proposition, to ascertain the addressee’s understanding, continued interest, or agreement, thus attaining or maintaining mutual alignment, to soften epistemic certainty and indicate politeness by avoiding possible face threats and assuring that the hearer is not feeling underestimated, to express triumph, as well as to signal that the proposition constitutes new information and is relevant for the thematic progression of the discourse. (81) a. Mark is not coming, see, he’s pretty busy with finals. b. It’s just that you are not impartial, see. c. See, I told you he would get the job! Nonetheless, a fact that contradicts the description made by the OED and at the same time reinforces another path of derivation, namely that it “likely derives from the imperative see” (Brinton 2008: 161), is that, despite these shared functions, there are some that are exclusive of see. Firstly, this DM can be used as a means to emphasize an order, reprimand, or objection by preceding it, thus, showing a harsher tone towards the addressee. Besides, when serving this function, it generally collocates with the adverbs now and/or here. Secondly, although it clearly has an intersubjective function and is focused on the addressee, it can also be found implying “a self-directed comment by the speaker confirming that he or she is getting the story straight (self-rather than other-confirming)” (Brinton 2008: 140). In this sense, it could behave as a 105 sequence adverb like “first.” Thirdly, it can also serve as a mere filler or hesitation marker, giving the speaker time to think or reflect before speaking. Lastly, there is a particularly interesting function which is linked to reported speech. The DM see was found as the first element in those instances of reported speech either as a quoted argumentation marker, i.e., maintaining one of its most frequent functions but as a verbatim replication, or marking the beginning of the quotation itself within the discourse. (82) a. See here, he was doing really well until he met you. b. When I went to the Isle of Man, see, I went on a boat and […] it only cost me ninepence. (BNC) c. There was…see…around 15 people there. d. And then he said, “see, that’s not my problem.” The next DM under this category is I see. In terms of its frequency in the three different corpora employed here, the searches yielded very low rates in general. The highest ones were found when occurring in sentence-final position, especially, in American English. Those were the only instances in which the number of occurrences per million words was higher than 1. This frequency is, then, far from the one attested in you see and even lower than see, which was already considered not that commonly used in British or Canadian English. The main function I see has is that as a backchanneling device (Scheibman, 2002: 98; Aijmer 2004: 264), i.e., “a short comment by the listener – like yeah, oh or mm-hu – to signal that he or she is paying attention and to encourage the speaker to go on” (Van Olmen 2010b: 231). This is a cooperative device extremely common in conversation which shows the hearer’s comprehension, interest, or agreement. It does, however, imply a stronger level of interest than other DMs with a similar function, such as yeah or okay (Biber et al 2021: 1086). Thus, it serves to attain or maintain alignment between the interlocutors at the same time as it manages turntaking during the communicative exchange. In a similar fashion to the previous two DMs, I see is related to the cognitive sense of see, ‘to understand’ and ‘to acknowledge.’ In fact, it keeps the same meaning both as a lexical verb and as a DM. It also marks politeness by assuring that the other participant is not being underestimated as well as by avoiding potential face threatening acts in those instances of (partial) disagreement. In those contexts, it is often followed by a but-clause and reflects a pragmatic mitigation strategy, since the speaker attenuates disagreement by showing comprehension and acknowledgment first. The way I see functions clearly contrasts with that of you see and see. They act as complementary parts in discourse. Whereas you see and see are attention-getting elements 106 which mark an explanation or justification for a prior claim, I see signals that attention is being paid as well as the agreement with or acknowledgment of the information provided by the speaker. Echoing Brinton’s (2008: 155-156) words: Thus, it could be said that I see functions not in parallel to, but in contrast to you see. While you see is a signal used by the speaker to indicate that an explanation or justification follows and to solicit the hearer’s acceptance or understanding of the explanation, I see is used by an addressee to denote acceptance of – or at least acknowledgment of – what the speaker has said. Nonetheless, that is not the only function I see serves. It is also used to express an opinion or observation based on reliable evidence and facts. In other words, it has the meaning of ‘as I can see’. The semantic implication concerning the source of the information is due to the evidential meaning present in see (refer to section 2.4.) and which is maintained in its discourse function. This way, despite the information being interpreted as subjective (hence, the pronoun ‘I’), it is also considered solid and trustworthy, thus strengthening the speaker’s epistemic stance. In certain contexts, it can also be perceived as having a harsher or defying tone, representing a face threat to the addressee. In these cases, the speaker is being ironical by saying that they agree with the addressee’s opinion when they actually think the opposite. That is to say, it signals that what the addressee thinks readily contradicts how the speaker perceives it. (83) a. That idea was already proposed here, I see, but I want to propose it again. b. I see, so you’ve thought it all out. c. Of course, everything I do is wrong, I see. In terms of the distribution of these three DMs, 107 they do not only show dialectal differences, but also different behaviors influenced by their corresponding functions. Therefore, a distinction could be made between you see and see on the one hand, and I see on the other. Whereas in British English you see and see appear predominantly in sentence-medial position (even though the difference in position is not as marked in see, with only a slight variation between the three types of distribution), in American and Canadian English they generally occur sentence-initial. Thus, they all tend to have a forward scope, i.e., they operate or encode information on the elements to their right, yet they differ in the length of that scope. In British, these DMs tend to operate only on part of the utterance rather than on the entire utterance, as it 107 Distributional frequency was obtained by searching instances of the DM in question preceded and followed by the corresponding punctuation mark. All genres present in the three corpora (i.e., all sources from which texts were obtained) were taken into account. These include spoken, fiction, magazines, newspapers, academic or miscellaneous. 107 is typical in American and Canadian English. It is also worth mentioning those instances in which they appear in sentence-final position. In such cases, they have a backward scope, i.e., they operate on the elements to their left, and normally serve and emphasize a more intersubjective function, for example, claiming mutual understanding and avoiding a potential face-threat to the addressee. This fact seems to extend to all DMs studied, not only to those within see-forms. With respect to I see, there is a clear tendency that extends throughout all three dialects, namely that it mainly appears on its own as an autonomous element making up a whole turn or in sentence-initial position. This is the expected result bearing in mind its main function as a backchannel response. Yet, when it expresses an opinion, i.e., the other function it also shows, it is usually placed in sentence-medial or sentence-final position. The rest of DMs within see-forms are not as frequent as the ones discussed above. These are as I see it, as/so you see, and let’s see together with its variants (let us see and let me see). The searches in the three corpora yielded a low rate of usage for all DMs regardless of their positions. The highest rates can be found in let’s see, which represents 1.92 instances per million words in American English and 0.81 in British English, both in sentence-initial position. Regarding their functions, as I see it is used as an attention-getting device. It introduces new information and manages turn-taking by signaling that the speaker takes the floor. It primarily functions as a marker to soften epistemic stance. By using this DM, the speaker is signaling that the utterance on which it operates represents a subjective opinion (hence, the first-person pronoun ‘I’), not a completely objective fact. Furthermore, there is even the possibility that the speaker may be wrong. The use of see, however, does imply that such an idea is based on somewhat reliable facts, as was also the case with I see, taking into account its evidential meaning. This provides certain relevance to the proposition for the progression of the discourse. In this respect, it can be found in those contexts in which the speaker does not want to disrespect the addressee’s opinion by nullifying it or imposing their own. Therefore, it is also marking politeness and serving as a hedge, since it provides a safe space for disagreement and does not involve a threat for the addressee’s negative face. (84) a. As I see it, we have a few problems here. b. We were good friends, as I see it. (COCA) c. The problem has, as I see it, nothing to do with that. As/so you see claims the addressee’s attention and expresses the speaker’s commitment to the truth value of the matrix clause, marking epistemic certainty. As you see is primarily used 108 with its original literal sense, as it claims the addressee’s attention not only towards the discourse, but also towards something that can be seen and is within their sight. It contrasts with as I see it, because it usually marks an objective fact that can be proven and observed by anyone rather than a subjective opinion. Conversely, so you see is more often used in a figurative and cognitive sense. It introduces a result or conclusion, so it could be paraphrased as ‘as you may conclude [from the preceding discourse].’ Therefore, it serves as a resultative conjunction. Apart from these specific functions, both variants share other usages which are in part related to the use of the second-person pronoun you. “By presenting information as if it were obvious, Speakers [sic] encourage its acceptance and minimize the need for supporting evidence” (Biber and Finegan 1988: 19). Consequently, they claim mutual understanding and seek to attain or maintain alignment between the interlocutors. Besides, they constitute a positive politeness strategy, since they establish common ground between the participants. (85) a. As you see, he has no experience at all in this field. b. The company, as you see, has refused pointing to the First Amendment concerns. (COCA) c. I am working, as you see. d. So you see, life isn’t always easy. e. I’ve heard that several times as well, so you see, maybe it’s real. Lastly, let’s see and its variants usually function as hesitation markers. They can be used “to indicate that one is thinking or trying to remember something before making a statement or answering a question” (OED, s.v. let me see in see (v.), sense P.2.b.ii). When the Object is the first-person-plural pronoun, us, it implies that both the speaker and the addressee are presented as “cooperators in recalling and presenting the information of the discourse” (Brinton 2008: 160). Hence, it serves to claim mutual alignment and to mark politeness. This implication is not present when the Object is the first-person-singular pronoun me. In that case, only the speaker is thinking or recalling information. These DMs can also be found in contexts in which the speaker is summarizing the previous information given by the addressee to assure that it has been understood correctly and claim the attention of the addressee. In such cases, it is also common for expressions along the lines of ‘right?’ or ‘is that correct?’ to appear at the end of the discourse, reinforcing that function of let’s see, let us see, or let me see. (86) a. Let’s see, there is one seat left on the next flight. b. He paid, let’s see, around $150. c. Let me see, the desk should be facing that direction. 109 d. This book was published in, let me see, 2015. e. Let’s see, so the suspect was already gone when you got home, is that correct? f. Let me see, this is not your first novel, right? In terms of distribution, a striking difference with respect to the rest of DMs here is that the predominant position remains the same in all three dialects, British, American, and Canadian English. Besides, these DMs follow the same pattern, that is, they are all largely found in sentence-initial position, having a forward scope. The only exception is recorded in Canadian English in the DM as I see it, which is more common in sentence-medial position. *Occurrences per million words Table 5. Distributional frequency of see-forms in PdE corpora. 3.3.2. Look-forms As attested with see, there are numerous DMs coming from the verb form look, namely look, look you, lookit, looky/lookee, and lookyhere/lookahere. Among these, the DM look has a significantly higher frequency. Not only that, but it is also extremely frequent in Present-day English in general when compared to other DMs from the physical perception field. It normally collocates with now and/or here and is typical of American English, showing a much lower frequency in British English. In fact, in the former it shows a rate of usage almost 15 times higher than in the latter. With respect to the rest of DMs here, they are not so common and show a frequency of less than 0.1 occurrences per million words. Look you is registered in British, American, and Canadian English, lookit is registered in British and American English, and looky/lookee and lookyhere/lookahere are restricted to American English —though there is only one instance of lookyhere recorded in the corpus. These data contrast with those presented by BNC COCA STRATHY Initial Medial Final Initial Medial Final Initial Medial Final you see 4.75 8.01 5.41 7.54 2.6 1.29 3.69 1.42 0.8 see 0.9 1.06 0.99 10.9 3.16 0.47 0.73 0.53 0.09 I see 4.15 0.95 0.76 1.98 0.06 0.74 5.22 0.07 0.47 as I see it 0.22 0.13 0.01 0.33 0.23 0.04 0.11 0.2 0.07 as you see 0.24 0.11 0.05 0.24 0.14 0.06 0.28 0.2 0.02 so you see 0.47 0.07 0 0.45 0.01 0 0.37 0.04 0 let’s see 0.81 0.16 0.08 1.92 0.17 0.01 0.46 0.16 0.02 let us see 0.02 0 0 0.02 0 0 0.09 0 0 let me see 0.48 0.21 0.01 0.51 0.01 0 0.31 0 0 116 within a larger sentence in the vast majority of cases. Likewise, in those few instances in which hear was recorded, it predominantly —not to say exclusively— appears in sentence-initial position. On the other hand, do you hear? and its variants all appear in sentence-final position, mostly having a backward scope except for those contexts expressed above and illustrated in (92c) in which it is used as a connective. The only DM within hear-forms that shows some variety is I hear. It mostly appears in sentence-final position in British and American English, whereas it is usually placed in sentence-medial position in Canadian English. It has to be noted, nonetheless, that except for American English, the results yielded for sentence-medial and final position only showed a slight difference between them, indicating that there is not a strong preference for any of those options. Thus, it just depends on whether the speaker wants to signal that only a particular segment of the proposition is subjective (sentence-medial) or wants to reinforce the subjective nature of the whole proposition (sentence-final). *Occurrences per million words Table 8. Distributional frequency of hear-forms in PdE corpora. 3.3.5. Feel-forms The only DM found in the semantic field of touching is I feel. It is not very frequent overall, as other equivalent DMs outside the field of perception are favored instead, such as I guess, I suppose, I believe, or I think. This DM does not assert the truth value of the utterance and chiefly functions as a means to attenuate the speaker’s epistemic stance towards the proposition by indicating that it represents a subjective opinion and should not be taken as an objective fact by any means. Actually, there is the possibility that the speaker may be wrong, reflecting a certain level of uncertainty. That is, it is similar to as I see it discussed above. Since it retains the same meaning as a lexical verb and as a DM, it bears the implication that such an opinion is based on feelings and emotions, rather than on actual evidence. It also follows the premise that speakers try to be BNC COCA STRATHY Initial Medial Final Initial Medial Final Initial Medial Final hear, hear 1.05 0 0 0.78 0 0 4.69 0 0 hear 0.01 0 0 0.01 0 0 0 0 0 I hear 0.01 0.07 0.08 0.03 0.06 0.19 0 0.11 0.09 do you hear? 0 0 0.43 0 0 0.16 0 0 0.22 you hear? 0 0 0.08 0 0 0.54 0 0 0.04 hear? 0 0 0 0 0 0.13 0 0 0.02 117 co-operative and polite in conversations, given that I feel is a politeness strategy that the speaker uses to assure that the addressee is not being underestimated and to avoid a potential threat to the addressee’s negative face. Therefore, in those contexts in which the opinion of the speaker may be too blunt or contradict that of the addressee, this DM serves as a mitigator, since it classifies the utterance as a mere opinion, without imposing it as an assertion, and leaves space for disagreement. (93) a. I feel, you are meant to have it. (COCA) b. You have, I feel, successfully portrayed your feelings in this painting. c. We have something to think about, I feel. Despite what it is seen in examples in (93), as a DM, I feel does not normally appear in sentence-initial position; such a spot, although possible, is quite infrequent. Hence, when it does, it is typically as a lexical verb followed by a that-clause. It is pre-eminently placed in sentence-medial position, having a forward scope, thus establishing that only that specific part of the proposition is subjective. On the contrary, when it is found in sentence-final position, it shows a backward scope, reinforcing the idea that the whole proposition is not a fact, but an opinion. *Occurrences per million words Table 9. Distributional frequency of I feel in PdE corpora. BNC COCA STRATHY Initial Medial Final Initial Medial Final Initial Medial Final I feel 0.06 0.47 0.07 0.06 0.21 0.07 0.02 0.18 0.04 118 119 CHAPTER 4. Strategies for semantic extension 4.1. Lexical ambiguity: Polysemy vs. Homonymy Before tackling the various strategies that lead to the semantic extension of physical perception verbs, the question of how to address this semantic extension, i.e., either as polysemous or homonymous forms, needs to be explored. As is well-known, a single word form, may it be a verb, an adjective, or a noun, can convey numberless different and distinct meanings, this is known as lexical ambiguity. 108 Because of this, utterances may have more than one interpretation, as that ambiguity can radically change the message. During any communicative exchange, the addressee has to disambiguate and attribute the appropriate interpretation by means of the linguistic and, especially, pragmatic context so that the message intended by the speaker can be properly understood and communication can be successful overall. Provided the ubiquitous presence of lexical ambiguity in language, such a lexical processing does not usually entail difficulties or much time for native speakers, it being almost automatic. In the case of foreign or second language learners, this could be more of a challenge, if they are not familiar with the meaning(s) in question. They would have to either ask the other participant or look it/them up in a dictionary and select the correct sense through contextual clues. A peculiar and interesting case is that concerning puns, for they are “the exception that illustrate the rule” Rodd (2018: 96-97). Humor has been proven to be quite an incredibly complex area of pragmatics even for advanced learners, since its comprehension requires not only linguistic but also sociocultural competence (Bell 2007: 28). Lexical ambiguity is the foundation on which puns are constructed and both meanings —not only the contextually appropriate one— must be known and accessed by the addressee for the humor of the pun to be understood. According to theoretical linguistics, the notion of lexical ambiguity can be subdivided into two main typologies which are also differentiated in dictionaries, namely polysemy and homonymy (Lyons 1977: 550; Panman 1982: 106; Klepousniotou 2002: 206; Klepousniotou et al 2012: 11; Bergenholtz & Agerbo 2014: 28). Despite these terms being widely used, their delimitation is not as clear in practice. The opposition between these concepts is traditionally based on whether the senses are semantically related, which defines whether there are one or 108 Some authors prefer the term ‘semantic ambiguity’ to refer to this same concept “as this makes it clear that it is the meaning of the word that is ambiguous and not its form or grammatical properties” (Vitello & Rodd 2015: 391). 120 more lexemes involved. Polysemy takes place between semantically related senses which usually have a common historical origin. Hence, one single lexeme is viewed as having several different related meanings. This is reflected in dictionaries by listing all those senses, which are usually numbered, within the same lexical entry. Homonymy, on the other hand, describes separate and semantically unrelated meanings. In this case, different meanings are expressed by different lexemes which just happen to share the same orthographic and phonological form. Therefore, they are represented in dictionaries as individual, separate lexical entries. To put it simply, polysemy shows a relation among the meanings, whereas homonymy is a mere accidental or coincidental phenomenon. POLYSEMY HOMONYMY Lexeme Meaning1 Lexeme1 Meaning1 Meaning2 Lexeme2 Meaning2 Meaningn Lexemen Meaningn Meaningn+1 Lexemen+1 Meaningn+1 Figure 5. Traditional distinction between polysemy and homonymy Although this could seem a simple and effective way to distinguish between them, there is a sizeable number of borderline cases in which such a division is not as easily applied. As a consequence, there have been several attempts to establish some definite and decisive criteria to differentiate between polysemy and homonymy, but none of them have provided an unequivocal answer in the end. Echoing Lyons’ (1977: 550) words, “[t]he difference between homonymy and polysemy is easier to explain in general terms than it is to define in terms of objective and operationally satisfactory criteria.” The two criteria which are conventionally used by linguists and lexicographers are (a) the etymological criterion and (b) the relatedness vs. unrelatedness of meaning criterion (Lyons 1977: 550-553). The etymological criterion analyses meaning from a diachronic perspective and states that lexemes with the same historical origin are instances of polysemy, given that the meanings evolve from the same lexeme and are thus related. Nonetheless, this is found insufficient and indecisive, since there are words whose origin is not certain even in well-recorded languages and it also “depends upon how far we are prepared to go […] in tracing the history of words” (Lyons 1977: 551). Moreover, native speakers are not always aware of the diachronic origin and development of words, which makes some authors question and even negate the relevance 121 of diachrony in synchronic analyses of languages (Lyons 1977: 551; Bergenholtz & Agerbo 2014: 29). The criterion concerning the relatedness vs. unrelatedness of meaning is not without problems either. It is indeed the one most commonly used as a basis to discriminate between polysemy and homonymy and highly correlates with the native speakers’ feelings. Yet, similar to the first criterion, this is not satisfactory or conclusive, as that (un)relatedness seems to be a matter of degree rather than an aspect accounted for in absolute terms, and in some cases native speakers’ intuitions do not coincide or are far from accurate. The property selection processes posited by Ibarretxe-Antuñano (1999: 162-178) could be helpful in discerning that relatedness, but it does not establish a core meaning and it is not clear how many of those properties should suffice to consider several meanings as related. Likewise, some authors have dealt with this issue by means of other theories, such as the family resemblance model (Taylor 2008: 49-51), what Lakoff (1987: 91-114) calls ‘radial categories’, or the compositionality approach (Neagu 2013: 332-334), facing similar results. Furthermore, homonymy has also been divided into two different notions, “partial homonymy” and “total” or “absolute homonymy” (Lyons 1977: 559-567). The first one refers to those cases in which one or more elements differ, the meaning and the morphological behavior and/or the syntactico-semantic one. The meaning, though different, can be said to be related to the prototypical or core meaning through any of the syntactico-semantic processes below or metaphorization, for instance (e.g., see1 meaning ‘to perceive with your eyes’ and see2 meaning ‘to understand’). If there is no relation whatsoever, that would be described as a case of total or absolute homonymy (e.g., bank1 meaning ‘the side of a river’ and bank2 meaning ‘the financial institution’). This unsolved situation has led to the fact that “[t]he classical view of polysemy and homonymy as a binary opposition does not prevail today as it did in the past” (Valera 2020). Instead of perceiving these two phenomena as a clear-cut dichotomy, there is a more widespread view which establishes a gradient or continuum between homonymy and polysemy (Cowie 1982: 51; Klepousniotou 2002: 206; Crossley et al 2010: 575; Murphy 2010: 91-93; 109 Klepousniotou et al 2012: 19; Valera 2020). These two phenomena, thus, represent the two ends of a continuum where items arrange themselves depending on their degree of semantic relatedness and their degree of entrenchment, i.e., how easily a semantic content is recalled and 109 Just in passing, both Crossley et al (2010: 575) and Murphy (2010: 91-93) add a third category in this continuum, vagueness, which describes lexical items with one general sense, but with many possible referents. It is placed at one of the ends of the gradient with polysemy staying in the middle. 122 activated in the mind (Murphy 2010: 92; Valera 2020). Partial homonymy is placed somewhere in the middle and total homonymy at one of the two ends of the continuum. As much as we agree with this view, it fails to address a fundamental question, at least for this study, how to treat this in lexicographical works, as the conventional distinction which reflects a clear dichotomy would not apply anymore. As mentioned above and earlier in section 2.2., in this study the distinction between homonymy and polysemy is not only accounted for in terms of meaning, but also according to the morphological, syntactic, and semantic behavior (i.e., the semantic frames). If these elements differ, it would be considered a case of homonymy and the different homonymous forms would be represented as different lexical entries in dictionaries. Although this could require much space in printed dictionaries which include all parts of speech, this can be easily done in online ones. Besides, the advantage of this approach is that the distinction of entries could be done more systematically, without having to purely rely on criteria based on intuitions. 4.2. Syntactico-semantic processes Among the strategies that lead to the creation of new homonymous forms, the syntacticosemantic processes that may take place cannot be left unnoticed. Their relevance can be explained by considering two main points: on the one hand, they represent the interconnection or interaction between the multiple syntactico-semantic distinct verbs conflated into a single verbal form; and on the other hand, they are “conducive not only to a variation in the completing features, but also, as a logical consequence, to different (sub)models of complementation and different semantic characterizations” (Comesaña Rincón et al 2013: 109). There are seven different processes that may be involved: (i) Causativity, (ii) Decausativization, (iii) Deletion, (iv) Insertion, (v) Promotion of Instrument to Subject, (vi) Reversibility, and (vii) Conativity. 110 Nevertheless, according to the data collected for the present study, physical perception verbs do not participate in all of them. The ones that have not been attested with these verbs are (ii) Decausativization, (v) Promotion of Instrument to Subject, and (vi) Reversibility, and so they will not be dealt with here. 110 Especially refer to Comesaña Rincón et al (2013: 107-144) and the references therein. 123 4.2.1. Causativity Causativity is a device through which a single verbal form can function according to two different syntactic models of complementation, a transitive and an intransitive one, creating two separate homonymous forms and, consequently, two separate lexicographical entries. Through this process the Subject of the intransitive construction becomes the Object of the transitive one, in which the causer of the process (hence, causativity) is introduced as the Subject. Interestingly, this takes place without the verbal unit losing its semantic content. The causer of the transitive construction can be of distinct semantic nature. In the case of physical perception verbs, it can be an Agent, a Force, or an Instrument, whereas the Object is normally an Affected. Accordingly, the Subject of the intransitive construction can have the role of Affected or Agent: (94) a. The alarm sounded when the enemies approached. b. The watchman sounded the alarm when the enemies approached. (95) a. The letter ‘e’ does not sound in ‘move.’ b. You don’t sound the letter ‘e’ in ‘move.’ This process superficially operates in parallel fashion to Decausativization, albeit with disparate and distinctive semantic and pragmatic implications regarding the intransitive construction. In the case of Causativity there is an assumed participant that is causing the process, though it is not overtly mentioned in the intransitive construction. Such an assumption, however, is not possible in Decausativization, since the construction does not in fact denote a process carried out by a virtual actor. 111 Therefore, there is no causer (they read the novel → the novel reads well). This has a consequence at the syntactic level. Whereas instances of Causativity can be transformed into equivalent passive constructions, those of Decausativization cannot, at least with the same semantic implication: (96) a. The alarm was sounded when the enemies approached (by the watchman). b. *The novel was read well (by them). 111 Following Comesaña Rincón et al (2013: 63), “ACTOR: It can be described as that entity that causes a PROCESS which, consequently, becomes an ACTION.” 124 4.2.2. Deletion Deletion is an utterly relevant and recurrent syntactico-semantic process. As its name suggests, it consists in the deletion of a subcategorized element of the verb chiefly due to pragmatic factors. 112 There are several subtypes depending on the specific element deleted: 113 (i) General Deletion of any kind of Complementation: it occurs when any completing structure (other than the Subject) is deleted for the loss of communicative relevance. Consequently, the verb changes its syntactic typology from a more complex one, such as transitive or adjunctive, to a simpler one, namely intransitive. This is illustrated in (97) and (98), where the Adverbial and the Object are respectively deleted: (97) a. Bob stared out of the window. b. I screamed and everyone stared. (98) a. I’m just watching their discussion. b. Are you going to help me with this or just watch? (ii) P → Ä by Deletion of PREP O: it occurs when the Object of a preposition is deleted, leading to a de-categorization of the non-contrastive preposition and its re-categorization as a non-contrastive adverb (following the terminology by Garrudo Carabias 1991, 1996). This way, the verb undergoes a change regarding its syntactic nature from a prepositional ditransitive verb, /SVOPO/, into a phrasal monotransitive verb, /SVOÄ/, as in (99): (99) a. His self-sacrifice saw him through his family crisis. b. I’m having a bad time, but I’m sure my friends will see me through. 114 (iii) Deletion of a Preposition: the last subtype involves the obligatory deletion of a preposition before a that-clause or a full infinitival construction. 115 This deletion does not create any repercussion on the semantic content of the verb as examples in (100) show. Particularly 112 Garrudo Carabias (1991, 1996) does view contextual deletion as a factor leading to the distinction of partial homonymous forms and, in its turn, of different lexical entries. Consequently, according to him, watch in are you going to play football or just watch? (Garrudo Carabias 1996: 779) and watch in We sat and watched the sunset would correspond to two separate entries, the former showing an intransitive pattern and the latter a transitive one. 113 See Comesaña Rincón et al (2013: 116 and ff) and the references therein. 114 Examples due to Garrudo Carabias (1996: 584). 115 There are no occurrences of a deletion of a preposition before a full infinitival construction in the physical perception verb class. Yet, they can be found in other verb classes: (vi) a. I mistook what he said for something unbelievable. b. *I mistook what he said for to mean that they were partners. c. I mistook what he said to mean that they were partners. (Garrudo Carabias 1996: 371) 125 interesting is the verb in (100c) since for the preposition not to be deleted, it allows the insertion of an anticipatory it, creating a separation between the preposition and the that-clause: (100) a. You must see that the doors are locked before you leave. b. *You must see to that the doors are locked before you leave. c. You must see to it that the doors are locked before you leave. 4.2.3. Insertion This process constitutes the structural counterpart of deletion and consists in the insertion (hence the name) of one element, namely a Complement, to a simpler pattern of complementation. As a consequence, the verb acquires new more complex syntactic patterns that affect its semantic content. In other words, when a verb’s syntactic complementation pattern goes from /SV/ to /SVC/ or from /SVO/ to /SVOC/, i.e., when there is an insertion of a Complement, a “new resulting state is usually added to the lexemic content of the verb” (Comesaña Rincón et al 2013: 134). 116 (101) a. Your feet smell. b. Your feet smell good. (102) a. We don’t want to see you. b. We don’t want to see you so unhappy. (Garrudo Carabias 1996: 582) Examples in (101) are particularly interesting. In spite of the fact that both verbs smell belong to the STATE parameter, they are neither syntactically nor semantically similar. To begin with, smell in (101a) behaves intransitively, in contrast with smell in (101b), which is a copulative verb requiring a Subject Complement. The insertion of this element triggers a specialization of meaning that makes the readings of both examples differ drastically, in fact, they are the complete opposite. The verb smell in (101a) does not mean ‘to have a smell,’ contrary to smell in (101b), but ‘to smell badly.’ 117 To put it in more linguistic terms, smell undergoes a specialization of meaning, modifying its inclusiveness by narrowing down its scope and becoming less general (Cruse 2020: 250). This way, it functions as a synonym of 116 In this case, we would be dealing with a small clause. Refer to footnote 76 as well as to section 3.2.2.3. of the present study. 117 Of course, if the inserted element functioning as Subject Complement had a negative connotation, say, bad, horrible, or disgusting, then such a drastic difference or opposition in the semantic readings would not take place. (vii) Your feet smell bad. = Your feet smell. 132 These properties unveil helpful linguistic information in the sense that they account for the reason why such mappings between each sensory modality and the domain of internal sensation take place. Unfortunately, they are not helpful enough to explain why these metaphorical associations are not systematically codified all verbal forms within the same field of physical perception, as will be shown in the ensuing sections. Hence, in spite of the fact that both Sweetser’s mappings and Ibarretxe-Antuñano’s properties offer undoubtedly valuable information regarding the metaphorization of physical perception, they still leave some questions unanswered. The reason behind this may lie in their approach to analyzing these associations. They deal with metaphorization through the analysis of sensory modalities and stop at that broad level rather than continuing with a deeper study on the specific verbal forms within each of them, leaving an open field for research. Although many scholars assert that perception metaphors are valid cross-linguistically and, hence, cross-culturally (e.g., Lakoff and Johnson (1980), or Sweetser (1990)), several anthropological and linguistic studies prove that they are not universal by any means. “[A]s a general tendency, cognitive linguists have overemphasized the universality of some of the metaphorical structures that they found and ignored the many cases of non-universality in metaphorical conceptualization” (Kövecses 2005: xii). The reason for such a situation is similar —if not equal— to the one concerning the hierarchy of the senses: it is a matter of ethnocentricity or, using Ibarretxe-Antuñano’s (2008: 25-26) words, “an omnipresent Western perspective that somehow ‘pollutes’ conceptual reality in the domain of perception.” The fact that these perception metaphors work for western societies and languages does not necessarily mean that they can be automatically extended to non-western communities without showing any type of variation. Regarding this point, Kövecses (2010: 88) argues that “[w]e cannot expect the exact same metaphors to occur in all languages, but we cannot expect metaphors that contradict universal human experience, either.” This claim seems quite reasonable bearing in mind that human beings are constrained by a shared physiology of the senses which makes perceptual experience to a large extent universal. Yet, that is not the only factor that motivates perception metaphors. We cannot —or, what is more, should not— ignore that how we use, understand, and, of course, conceptualize the senses is utterly subjective and culture-specific (Majid & Levinson 2011: 79). Indeed, how we “convey non-sensory experiences through sensory metaphors, [sic] is revealing not only of how we process and organize sensory data, but also of the sensory underpinnings of our culture” (Classen 2019: 26). 133 The notion of culture has often been overlooked in metaphor research, as this area overemphasizes the role of sensorimotor and physical experiences and, hence, the bodily-based grounding as the exclusive motivation for metaphorical thinking. Although the concept of embodiment originally comprised the socio-cultural dimension as well (Lakoff & Johnson 1980: 57), it gradually faded away from conceptual metaphor theory. It has not been until fairly recently that more interest has been raised on it (Kövecses 2005, 2010; Ibarretxe-Antuñano 2013a, 2013b; Caballero & Ibarretxe-Antuñano 2014) and more studies on conceptual metaphors in non-Western cultures have been put forward, highlighting cross-cultural differences and bringing back to the foreground the relevance of this component as key in the grounding of perception metaphors. A notorious and enlightening case in point is the mapping established between the domains of sight and cognition. Despite being described as universal on many occasions, for it is often found in European languages and is considered a primary metaphor (Grady 1997: 27), 125 there is ample and conclusive evidence that invalidates such a generally accepted assumption and, in turn, the universality of perceptual metaphors as a whole. Evans and Wilkins (2000) demonstrate that Australian aboriginal languages regularly recourse to the sense of hearing rather than sight as the source domain for the metaphorization of cognitive concepts. Seeger (1975: 213) also reports that in Suya, an indigenous language in Brazil, the same verbal form, ku-mba, is used to denote ‘to hear,’ ‘to understand,’ and ‘to know.’ Likewise, Mayer (1982: 246) shows that in Ommura, “[a]s elsewhere in Papua New Guinea, intellectual processes, knowledge and memory are associated with the ear” and, consequently, the verbal form iero means ‘to hear,’ ‘to know,’ and ‘to understand.’ Apart from hearing, other sense modalities are also attested as alternatives for the metaphorical expression of cognition. To name a few, the sense of smell is preferred in Luwo, a Western Nilotic language of South Sudan (Storch 2013: 59), and heat (i.e., touch) is used in Tzotzil, an indigenous Maya language in Mexico (Classen 1993, as quoted in Ibarretxe-Antuñano 2008: 25). All in all, inasmuch as these metaphorical mappings constitute a remarkable insight into the behavior and nature of this verb class and laid the foundation for innumerable studies — this included—, they show a rather important drawback: they only represent the general view and do not reflect the full metaphorical potential of physical perception verbs as source domains. Their metaphorical scope (as understood by Kövecses (2003: 80)) is, in fact, far more 125 Primary metaphors “are simple patterns […] which map fundamental perceptual concepts onto equally fundamental but not directly perceptual ones” (Grady 2010: 192). 134 extensive and richer than the one described above, which merely stands for a simplified analysis and overlooks numerous of their most common metaphorical schemas. To grasp their wide diversity, the different metaphorical extensions of each of the five main fields of perception will be explored. These meanings were chiefly collected from different dictionaries, both monolingual and bilingual. As will be seen below, there are cases in which some of these meanings are coded by several verbal forms of the same —or even a different— field whereas others are exclusive of a particular one. Therefore, in order to avoid generalizations, the conventionalized formula to express these mappings (i.e., ‘[MEANING] IS SEEING/HEARING/SMELLING/TASTING/TOUCHING’) has not been used here. 4.3.1. Sight Out of the five semantic fields of perception, sight is by far the most researched one and this, of course, applies to its metaphorical extensions too. As will be seen here, sight shows the highest and most varied number of metaphorical meanings. Its metaphorical scope revolves chiefly around cognition or, as called in Figure 6., knowledge. The association between these two domains is not unexpected. It is a well-known tendency present in western languages to use visual terms to convey mental processes, since sight is the “primary source of objective data about the world” and is also described as “the strongest and most reliable” sense (Sweetser 1990: 39). The meanings within this domain include ‘to understand’ (105a), ‘to imagine’ (105b), ‘to visualize’ (105c), ‘to foresee’ (105d), ‘to meditate, to think about’ (105e-f), ‘to regard, to consider’ (105g-h), ‘to examine’ (105i), ‘to realize’ (105j), ‘to know about somebody’ (105k), or ‘to plan’ (105l): (105) a. I see the point of your argument. b. I can’t see him winning, can you? c. I still see her as she was years ago. d. We all can see what will happen if she finds out. e. –Will you buy me that handbag? –I don’t know, I’ll have to see. f. We will look at the possibility of merging the two companies. g. She has always seen him as a little brother. h. We view things differently. i. I’ll have to see the contract first. j. I could see that he was trying to fool us. k. I haven’t seen much of him lately. 135 l. They look to start their own business in the future. As pointed out above, this field is not exclusively related to mental activity. There are some instances in which sight is metaphorically associated with other domains as well, specifically, (a) with social interaction, (b) with reliability and assurance, and (c) with internal feelings. Within the former domain, the following meanings can be found: ‘to meet’ (106a), ‘to meet by chance’ (106b), ‘to go out with’ (106c), ‘to receive a visit’ (106d), ‘to visit’ (106e), and ‘to accompany, to escort’ (106f): (106) a. I’m seeing him on Monday morning. b. Jake saw Zoe on his way back while he was driving. c. They’ve been seeing each other for two months. d. She’s too sick to see anyone these days. e. They went to see their grandparents last weekend. f. –Let me see you to the door. –Don’t worry, I’ll see myself out. Regarding these meanings, there are some aspects that are worth noting. In the first place, in examples (106a) and (106c), see appears in the progressive structure and both sentences are totally well-formed. This is due to an already mentioned fact. Even though the core or prototypical meaning of see rejects the progressive structure, the metaphorical meanings conveyed in these examples are associated with verbs and domains that admit it. Consequently, they inherently accept it in their own grammar and are readily expressed by this structural behavior. In the second place, when see means ‘to meet by chance,’ the verb lacks the syntactic and semantic reciprocity present in the meaning ‘to meet,’ which can also be conveyed by see iff accompanied by the reciprocal pronoun each other. 126 On this account, it is possible to restate (106a) as we are seeing each other on Monday morning, but it would be at least questionable to restate (106b) as ?they saw each other on his way back while he was driving. Despite such a sentence being syntactically and semantically well-formed, it cannot be viewed as having the same exact meaning as Jake saw Zoe on his way back while he was driving, for this does not 126 A word of caution is advisable here. The verb see meaning ‘to meet’ resembles a symmetrical verb, as it entails a logically symmetrical relationship and admits the reversibility or permutability of its arguments (i.e., if Jack is seeing (= meet) Sally on Tuesday, it also means that Sally is seeing (=meet) Jack on Tuesday). Nevertheless, this verb cannot be viewed as symmetrical itself. Contrarily to meet, which is inherently symmetrical, see necessitates the symmetrical proform each other to acquire such symmetry and admit the association of both participants in the Subject slot. Mind the contrast between the intransitive constructions Jack and Sally met and *Jack and Sally saw. Therefore, semantically speaking, see has the definitory feature of symmetrical elements, reversibility, albeit it is not syntactically formalized in its intransitive pattern. There are other lexical items which also behave akin to see, for instance, the verbs resemble or border, or the adjective like. 136 imply that Zoe saw him by chance too, i.e., there is no semantic reciprocity. Contrariwise, were we dealing with a context in which the two people saw each other by chance, that restatement would be well-formed and accurate. And in the third place, the meaning ‘to visit’ reflected in (106e) can acquire a further nuance depending on the Object referent. When the person visited is a specialist in the field, see also has the connotation of ‘to consult’ (107), as you are visiting that person with the purpose of consulting something or seeking for professional advice: (107) You should see a doctor before that rash gets worse. In (b), that is, the domain that has to do with reliability and assurance, the most representative meanings are ‘to ascertain, to find out’ (108a), ‘to make sure’ (108b), ‘to take care’ (108c-d) and ‘to be careful, to pay attention’ (108e): (108) a. Could you go and see who’s knocking at the door? (Garrudo Carabias 1996: 581) b. See that the kids are in bed by 8 o’clock. c. Could you watch the baby while we’re out? d. He was looking after some of his animals when the lightning struck. 127 e. Watch the cars! The last domain, (c), is that of internal feelings. In spite of the fact that this domain is prototypically related to the field of touch (see Figure 6.), there are some meanings that are expressed by means of sight verbs, namely ‘to experience, to suffer’ (109a), ‘to tolerate’(109b), ‘to admire, to respect’(109c), and ‘to consider inferior’(109d): (109) a. She has seen a great deal of pain in her short life. b. He can’t see a woman crying. c. She always looked up to her grandmother. d. He looked down on his co-workers. The meaning in (109a) represents a noteworthy case, because not only does it belong to a domain prototypically linked to a different field of physical perception, but also see can be readily replaced by the verb feel (She has felt a great deal of pain in her short life). Such a 127 An instance of diatopic variation has been recorded here. This meaning can also be conveyed by the prepositional verb see after in American English. Then, both options, the one with see and the one with look, can be used interchangeably in this dialect: (viii) He was looking/seeing after some of his animals when the lightning struck. 137 behavior can be explained on the grounds of Viberg’s (1984) work. As mentioned earlier in footnote 77, apart from establishing the well-known hierarchy of sense modalities (refer to Figure 1.), he (1984: 136-137) also noted that “a verb having a basic meaning belonging to a sense modality higher (to the left) in the hierarchy can get an extended meaning that covers some (or all) of the sense modalities lower in the hierarchy.” Those are not the only cases in which a domain or a meaning usually associated with another field of physical perception can be expressed by the field of sight. The meaning ‘to obey’ is traditionally metaphorically mapped onto the semantic field of hearing (refer to Figure 6. above), but the verb observe has been found denoting this meaning: (110) We must observe the rules. Lastly, there is a sizeable number of other senses that are created through metaphorization but, contrary to the previous ones, correspond to isolated meanings that are not semantically related among them. For this reason, they are all grouped together under the category ‘miscellany.’ Some of these meanings are: ‘to witness’ (111a), ‘to refer to’ (111b), ‘to seem’ (111c), ‘to last’ (111d), ‘to comment’ (111e), ‘to spy’ (111f), ‘to search’ (111g-h), ‘to visit a place’ (111i), ‘to face a particular direction’ (111j), ‘to read (about)’ (111k), and ‘to celebrate’ (111l): (111) a. This stadium has seen many thrilling football games. 128 b. See chapter 12 for further discussion. c. Your plan looks good! d. $500 should see you to the end of the month. e. ‘It wasn’t exactly like that.’ He observed. f. I think I’m being watched. g. I’ve looked everywhere, but I cannot find it! h. What are you looking for? i. I have always wanted to see South Korea. j. My office looks North. k. Her mother saw in the newspaper that Queen Elizabeth II passed away. l. Do they observe Christmas? 128 This example is the same as example (15b) presented in section 2.3.1. Nevertheless, its number has been changed for the sake of descriptive clarity and consistency in the present study. 138 Out of these meanings, ‘to visit a place’ is similar to ‘to visit’ from the social interaction domain. The difference lies in the Object referent that the particular verb requires. In the case of the social interaction domain, the verb requires a person, whereas in this case, the verb requires a place. 4.3.2. Hearing After sight, hearing occupies the second top position in the hierarchy and is “the sense of linguistic communication” par excellence (Ibarretxe-Antuñano 1999: 64; 2002: 101), which makes its appearance in everyday life extremely frequent. Metaphorically, it is primarily linked to heed, so the nuance ‘with attention’ is added to the prototypical meaning of hearing, thus creating ‘to pay attention.’ This semantic extension can be expressed with both listen and hear, yet the last one only admits it in affirmative constructions: (112) a. Do not listen to what other people say. b. You’d better hear what he has to say. c. *Do not hear what he has to say. Looking at Figure 6., it can be appreciated that this is not the only metaphorical meaning that can be found in the field of hearing. The meaning of obedience also appears as a metaphorical mapping, though this one is presented as a further development of heed. In other words, this meaning represents a step forward in the sense that it does not only denote a demand for attention, but also to do what is being said, that is, ‘to obey.’ Contrary to what happened with ‘to heed,’ only listen can convey this meaning: (113) She never listens to advice and, then, always regrets it. Interestingly, the field of hearing is also metaphorically associated with the domain of cognition, the prototypical mapping of sight. It should not be forgotten that hearing is our sense of linguistic communication and, thus, an important tool for exchanging information. It is in charge of speech reception, both external and internal, i.e., of hearing in its most strict sense as well as of intellectually processing what has been heard. Moreover, Sweetser (1990: 41) claims that hearing “is also our major means of intellectual and emotional influence on each other.” Thence, it is not far-fetched that this semantic field can also be related to cognitive processes. Within this domain, three extended meanings regularly used in standard communication are found: ‘to be told, to be informed’ (114a), ‘to understand’ (114b), and ‘to imagine’ (114c): 139 (114) a. I’ve heard that you found a new job. b. If I heard well, this is the correct answer. c. He could still hear her voice. Even though the first meaning does not belong to the cognitive domain strictly speaking, it is normally understood as ‘to know,’ because “to hear something is to become aware of it” (Rojo & Valenzuela 2005: 231). That is why it has been included in this group. In fact, we are dealing with an evidential meaning of hear (refer to section 2.4.) through which two pieces of information are being provided. On the one hand, we are implying that we know something, namely the content of the proposition, and, on the other hand, we are specifying that the source of that information is hearsay. The degree of reliability, however, is lower than if such a meaning were expressed by see. Regarding the second meaning, ‘to understand,’ it has been claimed that the presence of the adverb well is required for it to be inferred (Ibarretxe-Antuñano 2002: 103). Nevertheless, there is an example among the dictionaries consulted that rejects such assertion outright. That example is (115) below and appears in MEDAL, both in its printed and online version. (115) Look, you’re just not hearing me, are you? Finally, there are other metaphorical extensions that do not correspond to one particular domain and are grouped under the category ‘miscellany,’ similarly to what happened with sight. Some of these meanings are ‘to try, to judge’ (116a), ‘to grant’ (116b), ‘to attend’ (116c), ‘to seem’ (116d), ‘to try to discover carefully and informally’ (116e), ‘to pronounce’ (116f), or ‘to auscultate’ (116g): (116) a. Which judge will hear the case? b. Our Lord heard our prayers. c. I heard this recital last week. d. That idea sounds brilliant! e. Could you sound the investors out about what they are going to vote? f. You don’t sound the letter ‘b’ in ‘plumber.’ g. The doctor sounded her chest. The meaning ‘to grant’ in (116b) can be understood as a second development of ‘to heed,’ one of the main mappings in the domain of hearing and exemplified in (112b) above. As happened with the primary development of heed, ‘to obey,’ there is an effect or further action 140 after paying attention. In this example, the Lord not only paid attention to their prayers, but also granted their pleas. 4.3.3. Smell The semantic field of smell is considered to have fewer and less deep metaphorical meanings and, so, it is traditionally regarded as weaker on this matter than the rest of semantic fields of physical perception (Caplan 1973: 275; Sweetser 1990: 43; Neagu 2013: 340). Nonetheless, a closer look at it demonstrates that this is far from accurate. Such a claim is mostly based on the —mistaken— assumption that the semantic field of smell only expresses two metaphorical meanings (refer to Sweetser (1990: 37), for instance) when, in fact, smell verbs are a source for more metaphorical extensions than those two, as we shall prove below. Whereas sight, hearing, taste, and touch are mapped onto major semantic domains, Sweetser (1990: 38) associates smell with the target domain of dislikeable feelings, which can be described as a subdomain within the broader domain of feelings. The same mapping is also supported by most authors, for example, Viberg (2008: 158), who links smell with the feeling of unpleasantness. The motivation for this association may well be related to —if not rooted in— two facts: (i) there are more smell words referring to foul smells than good ones (Classen 2019: 20) and (ii) when smellST behaves intransitively, it undergoes a specialization of its meaning to convey ‘to smell bad,’ showing this way a similar meaning to stink. And these, in turn, may be partly due to our being exposed to foul smells more often than to foul tastes, for instance, given that smell is a distant sense and taste is a close one. To put it simply, whereas we can decide what to eat or drink, it is more difficult to do so with smells, because we do not need to have physical contact with that sort of stimuli as it is airborne (Classen 2019: 20). Two of the most frequent metaphorical meanings we can find in the smell field are ‘to suspect’ (117a) and ‘to guess, to sense something intuitively’ (117b-c). These two meanings may be similar, but there is an aspect that sharply differentiates them: though the first one is always used with a negative connotation, uses of the second one with positive implications are also possible, as (117d) illustrates: (117) a. If we don’t follow the plan, he will smell it. b. I smell trouble. c. They are able to smell out any corruption. d. He can smell a hit just by hearing a song once. 141 Regarding the first meaning, an interesting contrast can be drawn between the field of sight and the one in question that highlights the differences between both sense modalities. As has been mentioned numerous times throughout this study, our sense of sight is regarded as objective and reliable, which explains how the verb see acquired or developed the extended meaning ‘to know.’ Conversely, given the more subjective nature of smell, it is not considered as reliable, that is why the verb smell shows the metaphorical meaning ‘to suspect’ instead, reflecting that degree of uncertainty. Besides, it has also been claimed that the sense of smell is not as biologically developed and is even an impoverished sense (McGann 2017: 3). This way, even though these two fields can convey cognitive meanings, their epistemic values vary in virtue of the qualities of each of the sense modalities. Due to the fact that dogs use the sense of smell to trail things, verbs belonging to this semantic field are as a rule used instead of the actual verb trail when referring to animals searching for something (118a). This means that these verbs have not only acquired a new physical meaning, but that they have also replaced the original verb in daily use. As a consequence of such a semantic extension, the verbs sniff and, at a lower degree, smell have undergone a further extension of their meanings, but this time metaphorical, and are also used to mean ‘to investigate’ when the Subject-referent denotes people and not animals (118b). For this metaphor to be clearly interpreted, the preposition around or about is needed as a compulsory element of the verb: (118) a. The dogs have been sniffing the ground looking for a clue. b. The police have been sniffing *(around) here before. Within this domain of unpleasant or dislikeable feelings, other metaphorical meanings have been attested, namely ‘to show contempt’ (119a), ‘to be of inferior quality’ (119b-c), ‘to be bad or dishonest’ (119d-e), ‘to seem’ (119f), ‘to smack of’ (119g), and ‘to have a large quantity of something’ (119h): (119) a. ‘It’s hardly what I’d call elegant’, she sniffed. b. The plot is not bad, but the script smells. c. The book is good, but the movie stinks. d. Something about this deal really smells. e. The whole justice system stinks. f. The story didn’t smell right. Everything was quite suspicious. g. The proposal smelt of fraud. h. They stink of money. 244 Hansen, H.M. (1998) Function of Discourse Particles: A Study with Special Reference to Spoken Standard French. Amsterdam/Philadelphia: John Benjamins Publishing Company. --- (2006) “A dynamic polysemy approach to the lexical semantics of discourse markers (with an exemplary analysis of French toujours).” In K. Fischer (ed.), Approaches to discourse particles. Oxford: Elsevier, 21–41. Hardman, M.J. (1986) “Data-Source Marking in the Jaqi Languages.” In W. Chafe & J. Nichols (eds.), Evidentiality: The Linguistic Coding of Epistemology. Norwood, NJ: Ablex, 113–136. Hartmann, R.R.K. (1982) “Das Zweisprachige Wörterbuch im Fremdsprachenerwerb,” Germanistische Linguistik 3(6/80), 73–86. Haspelmath, M. & S.M. Michaelis (2017) “Analytic and Synthetic: Typological Change in Varieties of European Languages.” In I. Buchstaller & B. Siebenhaar (eds.), Language Variation – European Perspectives VI: Selected Papers from the 8th International Conference on Language Variation in Europe (ICLaVE 8), Leipzig, May 2015. Amsterdam: John Benjamins Publishing Company, 3–22. Hatherall, G. (1984) “Studying Dictionary Use: Some Findings and Proposals.” In R.K.K. Hartmann (ed.), LEXetcr83 Proceedings. Tübingen: Max Niemeyer Verlag, 183–189. Heine, B. (2013) “On Discourse Markers: Grammaticalization, Pragmaticalization, or Something Else?” Linguistics 51(6), 1205–1247. --- (2018) “Are there two different ways of approaching grammaticalization?” In S. Hancil, T. Breban, and J. Vicente Lozano (eds.), New Trends on Grammaticalization and Language Change. Amsterdam: John Benjamins Publishing Company, 23–54. Heine, B. & G. Kaltenböck (2021) “From clause to discourse marker: on the development of comment clauses,” Language Sciences 87, 1–16. Herbst, T. (1987) “A Proposal for a Valency Dictionary of English.” In R.F. Ilson (ed.), A Spectrum of Lexicography. Amsterdam: John Benjamins Publishing Company, 29–47. --- (1988) “A Valency Model for Nouns in English,” Journal of Linguistics 24, 265–301. 245 Herbst, T., D. Heath, Ian F. Roe, & D. Götz (2004) A Valency Dictionary of English: A Corpus Based Analysis of the Complementation Patterns of English Verbs, Nouns and Adjectives. Berlin, New York: Mouton de Gruyter. Hoenigswald, H.M. (1962) “Lexicography and Grammar.” In F.W. Householder (ed.), Problems in Lexicography. Bloomington: Indiana University Press, 103–110. Hopper, P.J. & E.C. Traugott (2003) Grammaticalization (2nd ed.). Cambridge: Cambridge University Press. Hornby, A.S. (1954) A Guide to Patterns and Usage in English. Oxford: Oxford University Press. --- (1975) Guide to Patterns and Usage in English (2nd ed.). Oxford: Oxford University Press. Horno Chéliz, Mª del C. (2002) “Aspecto Léxico y Verbos de Percepción. A Propósito de Ver y Mirar,” Archivo de Filología Aragonesa 59-60, 555–575. Howes, D. (1991) (ed.) The varieties of sensory experience: A sourcebook in the anthropology of the senses. (Anthropological horizons.) Toronto: University of Toronto Press. Huddleston, R. (1975) “Homonymy in the English Verbal Paradigm,” Lingua 37, 151–176. --- (1984) Introduction to the Grammar of English. Cambridge: Cambridge University Press. Huddleston, R. & G.K. Pullum (2012) The Cambridge Grammar of the English Language (5th ed.). Cambridge: Cambridge University Press. Ibarretxe-Antuñano, I. (1996) “Semantic extensions in the sense of smell,” Anuario del Seminario de Filología Vasca ‘Julio de Urquijo’ XXX(2), 631–643. --- (1999) Polysemy and Metaphor in Perception Verbs: A Cross-Linguistic Study (Unpublished Doctoral Dissertation). Edinburgh: University of Edinburgh. --- (2002) “MIND-AS-BODY as a Cross-Linguistic Conceptual Metaphor”, Miscelánea. A Journal of English and American Studies 25, 93–119. --- (2008) “Vision Metaphors for the Intellect: Are They really Cross-linguistic?” Atlantis 30(1), 15–33. --- (2013a) “The Power of the Senses and the Role of Culture in Metaphor and Language.” In R. Caballero & J.E. Díaz Vera (eds.), Sensuous Cognition. Explorations into Human Sentience: Imagination, (E)motion and Perception. Berlin/Boston: Mouton de Gruyter, 109–133. 246 --- (2013b) “The relationship between conceptual metaphor and culture,” Intercultural Pragmatics 10(2), 315–339. --- (2019) “Perception metaphors in cognitive linguistics.” In L.J. Speed, C. O’Meara, L. San Roque & A. Majid (eds.), Perception Metaphors. Amsterdam/Philadelphia: John Benjamins Publishing Company, 43–64. Ifantidou, E. (2001) Evidentials and Relevance (Pragmatics and Beyond New Series 86). Amsterdam/Philadelphia: John Benjamins Publishing Company. Izquierdo Alegría, D., R. González Ruiz, & Ó. Loureda Lamas (2016) “Un Acercamiento a los Fundamentos de la Evidencialidad y a su Recepción y Tratamiento en la Lingüística Hispánica.” In R. Ruiz, D. Alegría & Ó. Lamas (eds.), La Evidencialidad en Español. Madrid/Frankfurt: Iberoamericana/Vervuert Verlagsgesellschaft, 9–46. Jackendoff, R. (1972) Semantic Interpretation in Generative Grammar. Cambridge, MA: MIT Press. Jackson, H. (1985) “Grammar in the Dictionary.” In R.F. Ilson (ed.), Dictionaries, Lexicography, and Language Learning. Oxford: Pergamon Press, 53–59. --- (2013) Lexicography. An Introduction (2nd ed.). London: Routledge. Jacobsen, W.H. Jr. (1986) “The Heterogeneity of Evidentials in Makah.” In Chafe, W. & J. Nichols (eds.), Evidentiality: The Linguistic Coding of Epistemology. Norwood, NJ: Ablex, 3–28. Johanson, L. (1996) “On Bulgarian and Turkish Indirectives.” In N. Boretzky, W. Enninger, & T. Stolz (eds.), Areale, Kontakte, Dialekte. Sprache und ihre Dynamik in mehrsprachigen Situationen. Bochum, Germany: Brockmeyer, 84–94. --- (1998) “The Structure of Turkic.” In L. Johanson & É.Á. Csató (eds.), The Turkic Languages. London: Routledge, 30–66. --- (2000) “Turkic Indirectives.” In L. Johanson & B. Utas (eds.), Evidentials: Turkic, Iranian and Neighbouring Languages. Berlin: Mouton de Gruyter, 61–88. Johanson, L. & B. Utas (2000) Evidentials: Turkic, Iranian and Neighbouring Languages. Berlin: Mouton de Gruyter. Jørgensen, E. (1990) “Verbs of physical perception used in progressive tenses,” English Studies 5, 439–444. 247 --- (1991) “The progressive tenses and the so-called ‘non-conclusive’ verbs,” English Studies 2, 173–182. Joseph, B.D. (2003a) “Evidentiality in Proto-Indo-European? Building a Case.” In K. JonesBley, M. Huld, A. Della Volpe, & M. Robbins Dexter, Proceedings of the Fourteenth Annual UCLA Indo-European Conference (Los Angeles, November 8-9, 2002). Journal of Indo-European Studies Monograph Series 47. Washington, DC: Institute for the Study of Man, 96–111. --- (2003b) “Evidentials: Summation, Questions, Prospects.” In A.Y. Aikhenvald and R.M.W. Dixon (eds.), Studies in Evidentiality. Amsterdam: John Benjamins Publishing Company, 307–327. Julich, N. (2019) “Why do we understand music as moving? The metaphorical basis of musical motion revisited.” In L.J. Speed, C. O’Meara, L. San Roque, & A. Majid (eds.), Perception Metaphors. Amsterdam/Philadelphia: John Benjamins Publishing Company, 165–184. Kersten, A. & L. Smith (2002) “Attention to Novel Objects during Verb Learning,” Child Development 73(1), 93–109. Kirby, S., M. Tamariz, H. Cornish, & K. Smith (2015) “Compression and Communication in the Cultural Evolution of Linguistic Structure,” Cognition 141, 87–102. Kittilä, S. (2020) “Folklore as an Evidential Category,” Folia Linguistica 54(3), 697–721. Klepousniotou, E. (2002) “The Processing of Lexical Ambiguity: Homonymy and Polysemy in the Mental Lexicon,” Brain and Language 81, 205–223. Klepousniotou, E., G.B. Pike, K. Steinhauer, & V. Gracco (2012) “Not All Ambiguous Words Are Created Equal: An EEG Investigation of Homonymy and Polysemy,” Brain and Language 123, 11–21. Kövecses, Z. (2000) Metaphor and Emotion: Language, Culture, and Body in Human Feeling. Cambridge/New York: Cambridge University Press. --- (2003) “The Scope of Metaphor.” In A. Barcelona (ed.), Metaphor and Metonymy at the Crossroads: A Cognitive Perspective. Berlin/New York: Mouton de Gruyter, 79–92. --- (2005) Metaphor in Culture: Universality and Variation. Cambridge: Cambridge University Press. 248 --- (2010) Metaphor: A Practical Introduction (2nd ed.). Oxford/New York: Oxford University Press. --- (2019) “Perception and Metaphor: The Case of Smell.” In In L.J. Speed, C. O’Meara, L. San Roque, & A. Majid (eds.), Perception Metaphors. Amsterdam/Philadelphia: John Benjamins Publishing Company, 327–346. Kreidler, C.W. (2014) Introducing English Semantics (2nd ed.), New York: Routledge. Lakoff, G. (1966) “Stative Adjectives and Verbs in English,” Mathematical Linguistics and Automatic Translation, Report No NSF-17, Section I. Cambridge, Massachusetts: Harvard Computation Laboratory, 1–16. --- (1987) Women, Fire and Dangerous Things. What Categories Reveal about the Mind. Chicago and London: Chicago University Press. Lakoff, G. & M. Johnson (1980) Metaphors We Live By. Chicago: University of Chicago Press. --- (1999) Philosophy in the Flesh: The Embodied Mind and its Challenge to Western Thought. New York: Basic Books. Lamiroy, B. & P. Swiggers (1991) “The Status of Imperatives as Discourse Signals.” In S. Fleischman & L.R. Waugh (eds.), Discourse-Pragmatics and the Verb: The Evidence from Romance. London: Routledge, 120–146. Langacker, R.W. (1987) Foundations of Cognitive Grammar I. Theoretical Prerequisites. Stanford, Ca.: Stanford University Press. --- (2006) “On the Continuous Debate about Discreteness,” Cognitive Linguistics 17(1), 107– 151. Lazard, G. (2001) “On the Grammaticalization of Evidentiality,” Journal of Pragmatics 33, 359–367. Lázaro Carreter, F. (1980) “Transformaciones Nominales y Diccionario.” In F. Lázaro Carreter (ed.), Estudios de Lingüística. Barcelona: Crítica, 73–81. Leech, G. (2004) Meaning and the English Verb (3rd ed.). London: Routledge. Leech, G. & J. Svartvik (2002) A Communicative Grammar of English (3rd ed.). London: Longman. Leech, G., P. Rayson & A. Wilson (2001) Word Frequencies in Written and Spoken English: Based on the British National Corpus. London: Longman. 249 Lehmann, C. (1995) Thoughts on Grammaticalization. Berlin: Language Science Press. Lemmens, M. & H. Wekker (1986) Grammar in English Learners’ Dictionaries. Tübingen: Max Niemeyer Verlag. --- (1991) “On the Relationship between Lexis and Grammar in English Learner’s Dictionaries,” International Journal of Lexicography 4(1), 1–14. Levin, B. (1991) “Building a Lexicon: The Contribution of Linguistics,” International Journal of Lexicography 4(3), 205–226. --- (1993) English Verb Classes and Alternations: A Preliminary Investigation. Chicago: University of Chicago Press. --- (2008) “Dative Verbs: A Crosslinguistic Perspective,” Lingvisticæ Investigationes 31, 285– 312. Levin, B. & M. Rappaport Hovav (1991) “Wiping the Slate Clean: A Lexical Semantic Exploration,” Cognition 41, 123–151. Levin, B. & T. Rapoport (1988) “Lexical Subordination,” Chicago Linguistic Society 24, 275– 289. Levinson, S. (2010) “Questions and Responses in Yélî Dnye, the Papuan Language of Rossel Island,” Journal of Pragmatics 42, 2741–2755. Levshina, N. (2022) Communicative Efficiency: Language structure and use. Cambridge: Cambridge University Press. Lew, R. (2002) “Questionnaires in Dictionary Use Research: A Reexamination.” In A. Braasch & C. Povlsen (eds.), Proceedings of the Tenth EURALEX International Congress, EURALEX 2002 (Vol. 1). Copenhagen: Center for Sprogteknologi, Copenhagen University, 267–271. Luckner, J.L. & C. Cooke (2010) “A Summary of the Vocabulary Research with Students Who Are Deaf or Hard of Hearing,” American Annals of the Deaf 155(1), 38–67. Lynott, D. & L. Connell (2009) “Modality exclusivity norms for 423 object properties,” Behavior Research Methods 41(2), 558–564. --- (2013) “Modality exclusivity norms for 400 nouns: The relationship between perceptual experience and surface word form,” Behavior Research Methods 45(2), 516–526. Lyons, J. (1977) Semantics. Vol. 2, Cambridge: Cambridge University Press. 250 Majid, A. & S.C. Levinson (2011) “The Senses in Language and Culture,” The Senses and Society 6(1), 5–18. Mandler, J. (1996) “Preverbal Representation and Language.” In P. Bloom, M.A. Peterson, L. Nadel, & M.F. Garrett (eds.), Language, Speech, and Communication. Language and Space. Cambridge, MA: MIT Press, 365–384. Marchello-Nizia, Ch. (2009) “Grammaticalisation et Pragmaticalisation des Connecteurs de Concession en Français: Cependant, Toutefois, Pourtant,” Revue Roumaine de Linguistique 54(1–2), 7–20. Markie, P. & M. Folescu (2021) “Rationalism vs. Empiricism,” In E.N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Fall 2021 ed.). https://plato.stanford.edu/archives/fall2021/entries/rationalism-empiricism/. Martín Mingorance, L. (1990) “Léxico y Sintaxis en la Gramática Funcional de S.C. Dik.” Actas de las I Jornadas de Lengua y Literatura Inglesa y Norteamericana. Logroño: Colegio Universitario de La Rioja, 121–138. Martín Zorraquino, M.A. (2008) “Los marcadores del discurso desde el punto de vista gramatical.” In M.A. Martín Zorraquino & E. Montolío (eds.), Los marcadores del discurso. Teoría y análisis. Madrid: Arco Libros, 19–53. Matlock, T. (1989) “Metaphor and the Grammaticalization of Evidentials.” Proceedings of the Fifteenth Annual Meeting of the Berkeley Linguistics Society, 215–225. Matthews, P.H. (1981) Syntax. Cambridge: Cambridge University Press. Mayer, J.R. (1982) “Body, Psyche and Society: Conceptions of Illness in Ommura, Eastern Highlands, Papua New Guinea,” Oceania 52(3), 240–260. McGann, J. (2017) “Poor Human Olfaction Is a 19th-Century Myth.” Science 356(6338). pii: eaam7263. Mithun, M. (1986) “Evidential Diachrony in Northern Iroquoian.” In W. Chafe & J. Nichols (eds.), Evidentiality: The Linguistic Coding of Epistemology. Norwood, NJ: Ablex, 89– 112. --- (2020) “Information-Structural Perspectives on Discourse Particles.” In P.Y. Modicom & O. Duplâtre (eds.), Information-Structural Perspectives on Discourse Particles. Amsterdam/Philadelphia: John Benjamins Publishing Company, 27–46. 251 Modicom, P.Y. & O. Duplâtre (2020) “What can information-structural categories tell us about discourse particles?” In P.Y. Modicom & O. Duplâtre (eds.), Information-Structural Perspectives on Discourse Particles. Amsterdam/Philadelphia: John Benjamins Publishing Company, 1–24. Mora Gutiérrez, J.P. (2001) “Directed Motion in English and Spanish,” Estudios de Lingüística del Español 11. [Available at http://elies.rediris.es/publicaciones.html]. Mugdan, J. (1989) “Grammar in Dictionaries of Languages for Special Purposes (LSP),” Hermes 3, 125–142. Müller, N., A. Nagels, & Ch. Kauschke (2022) “Metaphorical Expressions Originating from Human Senses: Psycholinguistic and Affective Norms for German Metaphors for Internal State Terms (MIST Database),” Behavior Research Methods 54, 365–377. Murphy, M.L. (2010) Lexical meaning. Cambridge: Cambridge University Press. Murray, D.E. (2000) “Protean Communication: The Language of Computer-Mediated Communication,” TESOL Quarterly 34(3), 397–421. Nation, I.S.P. (2001) Learning Vocabulary in Another Language. Cambridge: Cambridge University Press. Neagu, M. (2013) “What Is Universal and What Is Language-Specific in the Polysemy of Perception Verbs?,” Revue Roumaine de Linguistique LVIII(3), 329–343. Nesi, H. (2013) “Researching Users and Uses of Dictionaries.” In H. Jackson (ed.), The Bloomsbury Companion to Lexicography. London: Bloomsbury, 62–74. Norcliffe, E. & A. Majid (2024) “Verbs of Perception: A Quantitative Typological Study,” Language 100(1), 81–123. Nuyts, J. (2001a) “Subjectivity as an Evidential Dimension in Epistemic Modal Expressions,” Journal of Pragmatics 33, 383–400. --- (2001b) Epistemic Modality, Language, and Conceptualization: A Cognitive-Pragmatic Perspective. Amsterdam: John Benjamins Publishing Company. O’Meara, C., L.J. Speed, L. San Roque, & A. Majid (2019) “Perception Metaphors: A View from Diversity.” In L.J. Speed, C. O’Meara, L. San Roque, & A. Majid (eds.), Perception Metaphors. Amsterdam/Philadelphia: John Benjamins Publishing Company, 1–16. 252 Palmer, F.R. (1986) Mood and Modality. Cambridge: Cambridge University Press. --- (1988) The English Verb (2nd ed.). London: Longman. Palmer, S.E. (1999) Vision Science: Photons to Phenomenology. Cambridge, MA: Bradford Books. Panman, O. (1982) “Homonymy and Polysemy,” Lingua 58, 105–136. Paul, P. (1982) “Homonyms, Semantic Divergence and Valency,” Lingua 58(3-4), 291–307. Pecher, D. & R.A. Zwaan (2005) “Introduction to Grounding Cognition: The Role of Perception and Action in Memory, Language, and Thinking.” In D. Pecher and R.A. Zwaan (eds.), Grounding Cognition: The Role of Perception and Action in Memory, Language, and Thinking. Cambridge: Cambridge University Press, 1–7. Pietrandrea, P. (2007) “The Grammatical Nature of Some Epistemic-Evidential Adverbs in Spoken Italian,” Rivista di Linguistica 19(1), 39–63. Pike, G., G. Edgar, & H. Edgar (2012) “Perception.” In N. Braisby and A. Gellatly (eds.), Cognitive Psicology. Oxford: Oxford University Press, 65–99. Poutsma, H. (1926) A Grammar of LateModern English. Part II: The Parts of Speech. Section II. Groningen: P. Nordhoff. Prévost, S. (2011) “A propos from Verbal Complement to Discourse Marker: A Case of Grammaticalization?” Linguistics 49(2), 391–413. Quirk, R., S. Greenbaum, G. Leech, & J. Svartvik (1972) A Grammar of Contemporary English. London: Longman. --- (1985) A Comprehensive Grammar of the English Language. London: Longman. Radford, A. (1988) Transformational Grammar. A First Course. Cambridge: Cambridge University Press. Ramos Cortés, M. (2006) “Gramática en los Diccionarios. Un Caso a Examen: Los Diccionarios de Aprendizaje del Inglés,” Odisea 7, 139–157. Riemer, N. (2010) Introducing Semantics. Cambridge: Cambridge University Press. Rizo Rodríguez, A.J. (1991) “Diccionarios de Aprendizaje Monolingües y Bilingües del Inglés: un Estudio Comparativo.” In I. McCandless (ed.), Silvia Nath: In Memoriam. Granada: Servicio de Publicaciones de la Universidad de Granada, 105–129. 253 --- (1992) “Los Códigos Gramaticales en los Diccionarios Monolingües Ingleses de Aprendizaje,” Revista de la Facultad de Humanidades de Jaén 1(1), 99–130. Rodd, J.M. (2018) “Lexical Ambiguity.” In M.G. Gaskell and S-A. Rueschemeyer (eds.), The Oxford Handbook of Psycholinguistics (2nd ed.). Oxford: Oxford University Press, 96– 117. Rodríguez Espiñeira, M.J. (2002) “Las Oposiciones Léxico-Gramaticales entre Mirar/Ver y Escuchar/Oír.” In Homenaxe a Fernando R. Tato Plaza. Santiago de Compostela: Universidade de Santiago de Compostela, 437–489. Rogers, A. (1971) “Three Kinds of Physical Perception Verbs,” Chicago Linguistic Society 7, 206–223. --- (1972) “Another Look at Flip Perception Verbs,” Chicago Linguistic Society 8, 303-315. Rojo, A.M. & J. Valenzuela (2005) “Verbs of Sensory Perception: An English-Spanish Comparison”, Languages in Contrast 5(2), 219–243. Romero Trillo, J. (1997) “Your attention, please: pragmatic mechanisms to obtain the addressee’s attention in English and Spanish conversations,” Journal of Pragmatics 28(2), 205–221. Rosch, E. (1973) “Natural Categories,” Cognitive Psychology 4, 328–350. --- (1975) “Cognitive Representations of Semantic Categories,” Journal of Experimental Psychology: General 104(3), 192–233. --- (1978) “Principles of Categorization.” In E. Rosch and B.B. Lloyd (eds.), Cognition and Categorization. Hillsdale, New Jersey: Lawrence Erlbaum Associates. Rosch, E. & C.B. Mervis (1975) “Family Resemblances: Studies in the Internal Structure of Categories,” Cognitive Psychology 7, 573–605. Roseano, P., M. González, J. Borràs-Comes, & P. Prieto (2016) “Communicating Epistemic Stance: How Speech and Gesture Patterns Reflect Epistemicity and Evidentiality,” Discourse Processes 53(3), 135–174. Rusiecki, J. (1985) Adjectives and Comparison in English: A Semantic Study. London: Longman. Saeed, A.T. & S. Fareh (2011) “Arab EFL Learners’ Acquisition of Verbs of Senses,” Studia Anglica Posnaniensia 47(1), 3–24. 260 Richards, J.C. & R. Schmidt (2010) (eds.) Longman Dictionary of Language Teaching and Applied Linguistics (4th ed.). London: Routledge. Rundell, M. & G. Fox (2007) (eds.) Macmillan English Dictionary for Advanced Learners (2nd ed.). Oxford: Macmillan. Sinclair, J. & R. Combley (2023) (eds.) Collins COBUILD: Advanced Learner’s Dictionary (10th ed.). London: Harper Collins Publishers. Summers, D. (2009) (ed.) Longman Dictionary of Contemporary English (5th ed.). Harlow Essex: Pearson. Online format Cambridge Advanced Learner’s Dictionary, http://dictionary.cambridge.org Collins COBUILD English Dictionary, https://www.collinsdictionary.com Diccionario de la Lengua Española, https://dle.rae.es Dictionary.com, https://www.dictionary.com/ LSD.Law, https://www.lsd.law/define/ Longman Dictionary of Contemporary English, https://www.ldoceonline.com Merriam-Webster Dictionary, https://www.merriam-webster.com Oxford Advanced Learner’s Dictionary, https://www.oxfordlearnersdictionaries.com Oxford English Dictionary, http://www.oed.com The Britannica Dictionary, https://www.britannica.com/dictionary The Free Dictionary, https://www.thefreedictionary.com 261 Wex (Legal Information Institute), https://www.law.cornell.edu/wex WordReference, https://www.wordreference.com Wordsmyth, https://www.wordsmyth.net