The heterogeneous nature of verbal alternations: What information structure can tell us about it

ISSN 2385-4138 (digital) Isogloss 2023, 9(4)/6 https://doi.org/10.5565/rev/isogloss.309 1-44 The heterogeneous nature of verbal alternations: What information structure can tell us about it Ángel L. Jiménez-Fernández University of Seville [email protected] Received: 03-03-2023 Accepted: 24-09-2023 Published: 15-12-2023 How to cite: Jiménez-Fernández, Ángel L. 2023. The heterogeneous nature of verbal alternations: What information structure can tell us about it. In Trending topics in Romance linguistics, eds Roberta Pires de Oliveira & Cilene Rodrigues. Special issue of Isogloss. Open Journal of Romance Linguistics 9(4)/6, 1-44. DOI: https://doi.org/10.5565/rev/isogloss.309 Abstract In this paper, I discuss Dative Alternation and Locative Alternation, proposing a cross-linguistically heterogeneous behavior of verb alternates. I explore information structural factors and study the available basic verb patterns in contexts where no previous shared information is presumed (broad focus), which is taken as diagnostic of the derived or non-derived nature of a given verbal alternation. I examine the different discourse interpretations of the two alternates in Spanish, English, Portuguese and Turkish. I suggest a non-derived nature of the two instances of verbs in Spanish Locative Alternation. On the contrary, Dative Alternation in Spanish exhibits derivational properties from a discourse perspective. In short, I propose two types of alternates (a derivational type and a non-derivational type), for which I find evidence in the notion of scope freezing (Larson 2014, Antonyuk 2020, Cépeda & Cyrino 2020). In Spanish, an inverse scope reading is available in quantifiers in locative alternation, which supports a non-derivational analysis of the two patterns. Locative Alternation Isogloss 2023, 9(4)/6 Ángel Jiménez-Fernández 2 shows scope freezing, favoring a derivational analysis. I put forth a Minimalist derivation based on discourse features and the notion of feature inheritance, which accounts for the above-mentioned characteristics. Keywords: verbal alternations, argument structure, information structure, broad focus, scope freezing. 1. Introduction Argument structure alternations have been a crucial area of research in generative linguistics since its beginning, yielding important implications for our understanding of the overall architecture of grammar per se as well as issues regarding the nature of the relation between lexicon and syntax in particular (Levin 1993, Levin & Rappaport Hovav 2005; see also Ramchand 2013 for a general overview). The examples in (1) illustrate the classic causative/inchoative alternation: (1) a. The chair broke. b. John broke the chair. c. The chair is broken. Examples in (1) raise the question as to whether we are dealing with the same verb – break– or with different verbs. In her paper on argument structure alternations, Ramchand (2013: 267) states that “distinctions in verb meaning must be encoded only insofar as they have systematic effects in the grammar”. Underlying this statement and the question of one vs. different verbs are the various approaches to verb alternations, namely derivational or non-derivational. In addition, approaches to a possible answer to this question have been based on the semantics of the two alternating constructions (Levin 1993, 2006) or on whether their syntax is different or similar (Larson 1988, 2014). Paying attention to the semantics of the alternates, Levin (2006) proposes a meticulous classification of verbs. In the search of a unified account, Levin (1993) produces an extensive spectrum of verb alternations, distinguishing between putting verbs (spray) and removing verbs (clear): (2) Locative Alternation — “putting” subtype: a. Jill sprayed paint on the wall. b. Jill sprayed the wall with paint. (3) Locative Alternation — “removing” subtype: a. Jack wiped crumbs off the counter. b. Jack wiped the counter. As we may observe, the crucial factor is the verb meaning (For Spanish, see Mayoral Hernández 2015). In addition to semantic elements, some classifications have considered syntactic aspects too. For instance, the preposition selected by each alternate in the pair or the syntactic category of the location argument. Once Mayoral Hernández The heterogeneous nature of verbal alternations Isogloss 2023, 9(4)/6 3 (2015) has established his semantic classification, he distinguishes different verbs in his first group for Spanish depending on the preposition involved: • Either con ‘with’ or de ‘of’, as in cargar ‘load’ (4) Marta cargó el coche con/de leña. Marta load-PST.3SG the car with of wood ‘Marta loaded her car with wood.’ • Only con ‘with’, as in cultivar ‘raise’ (5) a. Juan cultivaba cebollas en su huerto. Juan raise-PST.3SG onions in his vegetable garden b. Juan cultivaba su huerto con cebollas. Juan raise-PST.3SG his vegetable garden with onions ‘Juan raised onions in the vegetable garden.’ • Only de ‘of’, as in plantar ‘grow’ (6) a. Plantaron pinos en el monte. plant-PST.3SG pinetrees in the mountain b. Plantaron el monte de pinos. plant-PST.3SG the mountain of pinetrees ‘They planted the mountain in pinetrees.’ The second syntactic aspect taken into consideration by Mayoral Hernández (2015) is the category of the location argument. The first criterion is whether the location argument can occur in a PP: (7) a. El viento barrió de cenizas nuestras amadas calles. the wind sweep-PST.3SG of ashes our beloved streets b. El viento barrió de nuestras amadas calles las cenizas. the wind sweep-PST.3SG from our beloved streets the ashes ‘The wind swept the ashes from our beloved streets.’ The second criterion is whether location cannot occur as PP, including in this group verbs such as absolver (‘absolve’), aliviar (‘soothe’), curar (‘cure’), despojar (‘deprive’), desvalijar (‘clean out’), disculpar (‘excuse’), exculpar (‘exonerate’), etc.: (8) a. Disculpó a Felipe de sus horrendos crímenes. exonerate-PST.3SG to Felipe of his terrible crimes b. Disculpo *{de Felipe} los horrendos crímenes. exonerate-PST.3SG of Felipe the terrible crimes ‘He exonerated Felipe of his terrible crimes.’ For Agenjo (2019), in the approach to alternating verbs, one crucial question has been on what property makes a verb alternating. There are many problematic Isogloss 2023, 9(4)/6 Ángel Jiménez-Fernández 4 issues to decide on whether a particular verb is alternating or not. Focusing on Spanish, these are summarized as follows (Agenjo 2019: 33): 1) Different classifications (Cifuentes Honrubia 2008 or Mayoral Hernández 2015) show many controversial cases. 2) Speakers do not fully agree with what constitutes an alternating verb, making experimental work biased. 3) Dictionaries are also incoherent with the use of alternating verbs. Agenjo’s (2019) experimental work shows that, in Spanish, none of the 39 alternating verbs tested achieve 100 % acceptability (reaching only 60 % - 70 %, at most). Figure 1. Acceptability of alternating verbs in Agenjo’s experiment As observed in Figure 1, typical verbs which occur in all inventories are ranked very lowly (empapar ‘soak up’; exception grabar ‘engrave’), whereas verbs which are hard to find in research as alternating are ranked very highly (tatuar ‘get tatooed’). Agenjo (2019) also analyses the treatment of alternate verbs in usage-based dictionaries: Diccionario del Español Actual (DEA), Diccionario del Estudiante (DE), Diccionario Estudio Salamanca (DESAL) and Diccionario Clave (DC). The author arrives at a classification of verbs based on their occurrence as alternates in the above-mentioned dictionaries (Agenjo 2019: 20): The heterogeneous nature of verbal alternations Isogloss 2023, 9(4)/6 5 Figure 2. Agenjo’s findings in dictionaries The main problem in this classification is that classical alternate verbs from the literature such as rociar ‘spray’ are not even mentioned as alternate verbs. However, very infrequent verbs such as imbuir ‘infuse’ appear in the four dictionaries. What is clear is that there has been a monumental proliferation of alternations and classifications thereof in the literature, and none of them are free from shortcomings and flaws. Ramchand (2013) makes a substantial compilation of the most popular verb alternations in the literature, tackling the problem raised with the polysemous analysis (two lexical entries) or monosemous analysis (one single entry) that we may entertain for alternate verbs. To obtain a full picture of the scenario emerging from the dichotomy between derived and non-derived analyses of alternates, following is a summary of the alternations examined here. The first group that Ramchand (2013) discusses is the one involving Dative Shift or Dative Alternation, as illustrated in (9), including verbs such as give, send, buy, etc. (9) a. John gave the book to Mary. b. John gave Mary the book. Verbs here may select the patterns /OPO/, where the goal can be instantiated as a to-PP (as in to Mary in (9a)), or /OO/, where the goal is turned into the primary object (Mary in (9b)). As an object, this goal allows for passivization, as in (10): (10) Mary was given the book. A second verb Alternation that Ramchand takes into account is Locative Alternation, illustrated in (11): Isogloss 2023, 9(4)/6 Ángel Jiménez-Fernández 6 (11) a. John smothered the toast with marmalade. b. John smothered marmalade on the toast. In the Locative Alternation, the verb may be followed by the location in object position and then by what Ramchand defines as the ‘located substance’ preceded by the preposition with, as in (11a). Alternatively, the located substance can function as the direct object and the location is realized by a PP. Other verbs included in this group are spread, spray (paint), etc. A third class of verb alternation that Ramchand discusses is the Contactive Alternation (also called the with/against alternation, by Levin 1993). As illustrated in (12a), the direct object can be the contacted element followed by the instrument as a with-PP. However, in (12b) the verb selects the instrument as the direct object and the contacted object is expressed as a locative PP (Ramchand 2013: 276): (12) a. John hit the table with the cricket bat. b. John hit the cricket bat against the table. Other verbs included in the class of Contactive Alternation are strike, slap (somebody on a body part), etc. A fourth class is the so-called Causative-Inchoative Alternation, illustrated in (13). (13) a. The window broke. b. John broke the window. Given that the number of arguments selected by each of the alternate verbs is different, it is extremely difficult to propose a non-derived analysis. Much on the contrary, it is widely claimed that the verb exhibits two separate lexical entries. Other verbs in the alternation include ring (the bell), sink (the boat), burp (the baby), etc. In the proliferation of alternates, a crucial problem lays in the fact that some verbs may enter different lexical classes (Levin 1993), and hence showcase different alternations. This is the case of slide: (14) a. Carla slid the book to Dale. b. Carla slid Dale the book. (15) a. Carla slid the books across the table. b. The books slid across the table. In (14) slide involves Dative Alternation, whereas in (15) it exemplifies Causative/lnchoative Alternation. We have seen that, in addressing verb alternations, linguists have observed syntactic factors, semantic factors and/or lexical factors. However, only a few studies have paid attention to discourse and information structure factors. Bresnan et al. (2007) analyze the connection between definiteness, givenness and ordering in Dative Alternation in English, based on internet searches for examples. Speakers are supposed to show preference for the order /OO/ or /OPO/ depending on whether the relevant postverbal constituent is definite and displays given information. This is just one factor The heterogeneous nature of verbal alternations Isogloss 2023, 9(4)/6 7 influencing the ordering chosen by native English speakers. The authors claim that pronominal constituents precede non-pronominal ones, definite constituents precede indefinite ones and shorter elements precede longer ones (Bresnan et al. 2007). Examples in (16) are illustrative: (16) a. ??Karen hand-carried a man a form. b. Karen spoke with Gretchen about the procedure for registering a complaint, and hand-carried her a form, but Gretchen never completed it. As is clear, (16b) complies with all the factors mentioned about pronominality, definiteness and length, contrary to (16b). I agree that choosing one pattern or another is multifactorial. However, Bresnan et al. (2007) do not take these information structural factors into account in showing whether a polysemous or monosemous analysis is more accurate for English Dative Alternations. See Krifka (2003) for the information structure view of the two possible analyses in English Dative Alternation. For Spanish, Jiménez-Fernández (2009) has shown that the original ordering in ditransitives in Spanish dative constructions is /OPO/ and that the reverse ordering is obtained by some discourse-based operation whereby the first object is seen as topic and the second object as focus, and proposes an analysis based on cartography (Rizzi 1997) to distinguish between the two patterns. Finally, for Brazilian Portuguese, Scher (1996) and Armelin (2011) discuss the information structure of Dative Alternations arguing for an adjunct topic position in charge of the rearrangement of the two patterns /DO + IO/ and /IO + DO/. This is illustrated for BP as follows: (17) a. O João deu o livro ao/pra Pedro. the João give-PST.3SG the book to.the/for.the Pedro b. O João deu ao/pra Pedro o livro. ‘John gave a book to Peter.’ (Armelin 2011: 16, her examples in (1)) IN BP, the preposition alternates between a ‘to’ and pra ‘for’. There may be no preposition at all, as in the BP variety in Zona da Mata Mineira. What is interesting is that the ordering in Dative Alternation in BP displays exactly the same discourse properties as Spanish, and hence the DP o livro is a topic in (17a) but focus in (17b). Also, both Scher (1996) and Armelin (2011) claim that for broad focus, the preferred option (and hence the canonical construction) is the /DO IO/ pattern, similar to what Cépeda & Cyrino (2020) and Lacerda (2017, 2020) propose. Note that the constructions all these authors study are the two alternates in Dative Alternation. In an experimental perspective, Quarezemin (2009) has shown that the topic+focus partition in BP headily draws on prosodic properties. A task-based experiment displays a critical preference in BP to mark focus by intonation, instead of syntactic rearrangement. Quarezemin adopts a cartographic view and puts forth a Topic Phrase in the low periphery to account for the possible rearrangement. A Topic Phrase is also assumed by Cépeda & Cyrino (2020). From this brief discussion on previous information structure approaches to ditransitives, several conclusions are drawn. First, linguists have tended to focus only on Dative Alternation -no discourse-based study has been found on Locative Alternation. Second, it is clear that information structure is involved in the derivational Isogloss 2023, 9(4)/6 Ángel Jiménez-Fernández 8 connection between the two patterns of Dative Alternation. All authors agree that there is a topic position in ditransitives, and also that the primary basic pattern is /DO IO/ and hence express broad focus. However, they do not discuss the possible association of this original word order and its role in the possible derivational (monosemous) or non-derivational (polysemous) nature of verbal alternations. To fill this gap, the main goal of this paper is to explore a discourse-based explanation for two types of verbal alternations, namely Dative Alternation and Locative Alternation, based on the behavior of a given pattern in broad focus scenarios. My research question is whether information structure (in particular, broad focus) supports a derivational or a non-derivational view of argument structure alternations. In answering the question, my starting hypothesis is that there are two types of alternations inter- and cross-linguistically, namely 1) purely lexical-syntactic alternations, and 2) discourse-based alternations. The organization of this paper is as follows: In Section 2, I address the two approaches to verbal alternations. I focus on two classes of alternations, namely Dative Shift and Locative alternation in English and Spanish. In Section 3, I propose that, when Information Structure is taken into account, a completely different perspective is obtained in which Dative alternation involves two patterns which are related derivationally, whereas Locative alternation verbs are not related and require two separate lexical entries. In other words, I claim that there are two types of alternations cross-linguistically and that there is no homogeneous analysis for all verbal alternations. I extend the analysis to other languages such as Portuguese (a language in which objects are marked very similar to English and Spanish) and Turkish (a typologically different language which marks objects by using Case and/or Prepositions). Section 4 provides evidence for the split into 2 groups: scope freezing in quantified expressions. In Section 5, I present a possible formal analysis of the heterogeneous classes of alternates. Finally, Section 6 presents the general conclusions to which I arrive. 2. Two approaches to verb alternations In discussing alternations such as (18) in Double Object Constructions, there have been two views. (18) a. John has sent a letter to Mary. b. John has sent Mary a letter. First, proponents of the monosemy/thematic paraphrases view often take the position that one of the alternating frames derives from the other (Fillmore 1968, Larson 1988, 1990, 2014, Aoun and Li 1989, Baker 1997). In this derivational approach, lexical entries share the same underlying structure. Second, proponents of the polysemy view adopt the stance that the two constructions encode different semantic relations – for example, change of possession for the DOC and movement to a goal for a PP Dative (Jackendoff 1990, Pesetsky 1995, Harley 1995, 2002, 2014). This is the non-derivational approach, where lexical entries involve different semantic relations, and hence require different underlying structures. The heterogeneous nature of verbal alternations Isogloss 2023, 9(4)/6 9 The two approaches use arguments which might work for the purposes of establishing a possible relation between alternating verbs or not. Three different arguments can illustrate this in the following manner: As pointed out by Goldberg (2002: 337-340), there is not always a one-to-one correspondence in the alternation. This is the case of the pair load vs. cover: (19) a. Pat loaded the hay onto the wagon. b. Pat loaded the wagon with hay. (20) a. They covered the wall with posters. b. *They covered posters onto the wall. While load allows the alternation in (19), cover occurs in the frame /O + with- PP/ in (20a) but blocks the alternation with the frame /O + locative PP/ in (20b). The very same constraint is found in Spanish: (21) a. Pedro cargó el heno en el vagón. Pedro load-PST.3SG the hay in the wagon ‘Pedro loaded the hay onto the wagon.’ b. Pedro cargó el vagón con heno. Pedro load-PST.3SG the wagon with hay ‘Pedro loaded the wagon with hay.’ (22) a. Cubrieron la pared con posters. cover-PST.3PL the wall with posters ‘They covered the wall with posters.’ b. *Cubrieron posters en la pared. cover-PST.3PL posters on the wall ‘*They covered posters onto the wall.’ Furthermore, there is not always a one-to-one correspondence in meaning. In one of the frames of the alternation, a holistic interpretation is detected, which is absent in the other frame (Antonyuk & Mykhaylyk 2022, Rappaport-Hovav & Levin 1988, Kearns 2011): (23) a. John loaded the cart with hay. b. John loaded hay in the cart. (23a) implies the full loading of the cart, whereas (23b) does not (Anderson 1971). Again, the same distinction is found in Spanish: (24) a. Juan cargó el carro con heno. Juan load-PST.3.SG the cart with hay ‘John loaded the cart with hay.’ b. Juan cargó heno en el carro. Juan load-PST.3.SG hay in the cart ‘John loaded hay in the cart.’ Isogloss 2023, 9(4)/6 Ángel Jiménez-Fernández 16 b. Juana cosió el disfraz con los lazos. (49) A: ¿A qué cosió Juana los lazos? ‘What did Juana sew with the bows?’ B: a. Juana cosió los lazos en el disfraz. b. # Juana cosió el disfraz con los lazos. 3.2. Extending the discourse-based view to Portuguese and Turkish In the previous section, I have shown that in locative alternation, the two patterns are plausible answers for a question such as ‘What’s up?’, which indicates that both qualify as all-focus sentences. This is taken as evidence supporting a polysemous analysis of (at least some) locative alternation verbs in Spanish and English. This observation can be extended to other languages such as Brazilian Portuguese, where a verb such as carregar ‘load’ exhibit two possible alternates, namely either a withmarked object for the locatum or a locative preposition with the location: 6 (50) a. O Higor carregou o caminhão com a lenha. the Igor load-PST.3SG the truck with the wood ‘Igor loaded the truck with the wood.’ b. O Higor carregou a lenha no caminhão. the Igor load-PST.3SG the wood in.the truck ‘Igor loaded the wood into the truck.’ Both can be answers to O que aconteceu? ‘what's up?’ (broad focus), supporting the view that this particular verb has two separate lexical entries. Another minimal pair is obtained with the verb encher ‘fill’, as illustrated in (51): (51) a. A Maria encheu a chaleira com água. the Maria fill-PST.3SG the teapot with water ‘Mary filled the teapot with water.’ b. A Maria encheu de água a chaleira. the Maria fill-PST.3SG of wáter the teapot ‘Mary filled water in the teapot.’ As is clear from (51), the verb encher ‘fill’ may be used in either pattern in allfocus sentences, giving credit to a polysemous analysis. Now I turn to Turkish locative alternation. Turkish is a language where all nominals are marked with morphological case, the default focus position is always immediately preceding the verb and where specificity effect arise associated with 6 An anonymous reviewer claims that there are other alternates for sentences such as (50b) which may sound more natural in BP (O Higor carregou com a lenha o caminhão). This is a very interesting point since it shows that, in BP, this particular verb may display an information structure-based alternation (where just word order changes) alongside a nonderived alternation. Also, notice that intonation is relevant in all pairs, given that a special pitch accent is used to indicate focus. The heterogeneous nature of verbal alternations Isogloss 2023, 9(4)/6 17 accusative case and syntactic position (Enç 1991). Once these factors are controlled, minimal pairs such as that in (52) may be obtained: (52) a. Can araba-yı odun-la doldur-du. Can.NOM cart-ACC wood-with fill-PST.3SG Lit. ‘Can filled the cart with wood.’ b. Can odun-lar-ı araba-ya doldur-du Can.NOM wood-PL-ACC cart-DAT load-PST.3SG ‘Can loaded wood on to the cart.’ Both patterns are possible in an out-of-the-blue context, which implies that both are basic patterns in Turkish. In (52b) we find the sequence /Acc + Dat/, where the locatum is marked with accusative and the location, with dative. In contrast, in (52a) the location is accusative whereas the locatum object is marked with a withpreposition. In short, locative alternation verbs such as doldur ‘fill’ can occur in either of their patterns as reply to a question such as What’s up?, which supports a nonderivational analysis of the verb. 3.3. Dative alternation: The Spanish/Portuguese case Two approaches can be found for the two variants in dative alternation in Spanish and Portuguese. The first approach suggests that the verbs in this group involve a double object construction (DOC) when doubled by a dative clitic, and the sequence O+PP- dative when there is no clitic (Demonte 1995, Cuervo 2003, Bleam 2003, among others): (53) a. Le entregué las llaves al conserje. CL-DAT.3.SG hand-PST.1SG the keys to.the janitor b. Entregué las llaves al conserje. ’I gave the keys to the janitor.’ (Demonte 1995: 6-7) Torres Morais & Salles 2010 argue for the same distinction between DOC and O+PP-dative in Brazilian Portuguese. 7 In this view, the two alternates have different lexical entries, hence a nonderivational analysis is proposed for the alternation. One potential problem that this approach poses is that not all varieties of Spanish showcase the optional insertion of the clitic, especially with most frequently used verbs such as dar ‘give’. In varieties such as Andalusian Spanish, the absence of the clitic yields a most unnatural outcome (NGRAE 2009): 7 DOC in Brazilian Portuguese is not a very productive syntactic process as the claim by Torres Morais & Salles would lead one to imagine, as an anonymous reviewer points out. This kind of structure is possible in some varieties of Brazilian Portuguese spoken in some states such as Minas Gerais, Goiás and Bahia. This issue is orthogonal to my main concern about the homogeneous or heterogeneous nature of ditransitives. For the reader interested in the presence or absence of the preposition in the IO in BP (in particular, in Minas Gerais), see Scher (1996) and Armelin (2011). Isogloss 2023, 9(4)/6 Ángel Jiménez-Fernández 18 (54) a. Le di las llaves a María. CL-DAT.3.SG give-PST.1SG the keys to Maria b. ?? Di las llaves a María. ‘I gave Maria the keys.’ In a different approach, the two variants in the alternation are derivationally connected. Here, word order plays a crucial role (Jiménez-Fernández 2009): (55) a. Le entregué las llaves al conserje. CL-DAT.3.SG hand-PST.1SG the keys to.the janitor b. Le entregué al conserje las llaves. ’I gave the keys to the janitor.’ Word order rearrangement is replicated in BP (Cépeda & Cyrino 2020): (56) a. A Olga deu uma maçã para o Mario. the Olga give-PST.3SG an apple to the Mario b. A Olga deu para o Mario uma maçã. ‘Olga gave an apple to Mario.’ Word order and information structure work hand in hand. From this perspective, the question arises as to whether there is any preference for a particular ordering in Spanish and Portuguese. Let us start with Narrow Focus. For Spanish, the final position is reserved for information focus. Therefore, depending on whether the indirect object or the direct object is focused, speakers will show a preference for /O+Dative-PP/ or /Dative-PP+O/, as illustrated in (57) and (58), respectively: (57) A: ¿A quién le diste las llaves? ‘Who did you give the keys?’ B: a. Le entregué las llaves al conserje. b. # Le entregué al conserje las llaves. ’I gave the keys to the janitor.’ (58) A: ¿Qué le diste al conserje? ‘What did you give to the genitor?’ B: a. # Le entregué las llaves al conserje. b. Le entregué al conserje las llaves. ’I gave the keys to the janitor.’ Cépeda & Cyrino (2020) also discuss the discourse-based choices on a particular sequence for ditransitives in BP when narrow focus is concerned, claiming that there is also a preference for a particular sequence. This does not mean that the sentences dubbed as # cannot ever be used as answers for the context question. They can, but they are not the preferred option in BP, at least for Cépeda & Cyrino: 8 8 In the last two decades, there has been a fruitful debate about how BP marks focus. Rearrangement of word order seems to be one possible way (Cépeda & Cyrino 2020). However, an anonymous reviewer states that the preference in BP is for focus in situ (Quarezemin 2009, Lacerda 2020), marked prosodically. Both answers in (53) are thus The heterogeneous nature of verbal alternations Isogloss 2023, 9(4)/6 19 (59) A: A quem a Olga deu uma maçã? ‘Who did Olga give an apple?’ B: a. A Olga deu uma maçã para o Mario. b. # A Olga deu para o Mario uma maçã. ‘Olga gave an apple to Mario.’ (60) A: O que a Olga deu para o Mario? ‘What did Olga give to Mario?’ B: a. # A Olga deu uma maçã para o Mario. b. A Olga deu para o Mario uma maçã. ‘Olga gave an apple to Mario.’ As for broad focus, there seems to be a difference between the two patterns. Only the sequence /O+PP-dative/ is optimally picked up as an all-focus sentence. (61- 62) illustrate the situation is Spanish; (63) does so in Portuguese: (61) A: ¿Qué pasó? ‘What happened?’ B: a. Le concedieron el Premio Nobel a Cela. CL-DAT.3SG award-PST.3PL the Prize Nobel to Cela b. # Le concedieron a Cela el Premio Nobel. ‘Cela was awarded the Nobel Prize.’ (adapted from Demonte 1995) (62) A: ¿Qué pasó? ‘What happened?’ B: a. Juan le enseñó el vídeo a Rosalía. Juan CL-DAT.3SG show-PST.1SG the video to Rosalia b. # Juan le enseñó a Rosalía el vídeo. ‘Juan showed the video to Rosalía.’ (63) A: O que aconteceu? ‘What happened?’ B: a. A Olga deu uma maçã para o Mario. b. # A Olga deu para o Mario uma maçã. ‘Olga gave an apple to Mario.’ The conclusion we can draw from Ibero-Romance dative alternation is that, as far as all-focus sentences (broad focus) are indicative of the canonical pattern in a given language (under the question-answer congruence; Krifka 2008; Repp 2016; Bentley & Cruschina 2018; Fábregas et al. 2016), only the order DO+PP-dative is attested as basic in Spanish and Portuguese (Jiménez-Fernández 2009 for Spanish; Lacerda 2020, 2017 for Portuguese, among many others). This is confirmed in narrow acceptable with the appropriate intonation, given that BP is not a focus-final language, whereas Spanish is. There are multiple strategies that languages may use to express information structure, two of them are prosody and syntactic reordering, which can be used in one single language, as is the case in Spanish (Zubizarreta 1998). What is relevant for the BP examples in (59-60) is that, syntactically speaking, the preference is always the one putting the constituent satisfying the question’s information request in final position. Prosodically, there may be other alternatives, which do not concern us here. Isogloss 2023, 9(4)/6 Ángel Jiménez-Fernández 20 focus sentences where the preference for one or other member of the alternation is crucially dependent on the discourse function of the DO and the PP-dative. One final note is in order regarding dative shift in Turkish. The order of the objects in double object constructions is in constant debate in Turkish (Issever 2003, Kornfilt 2003, Simpson et al. 2008, Georgala 2011). For linguists such as Kornfilt 2003, the basic order is /DO + IO/, whereas for others such as Georgala (2011), the canonical order is /IO + DO/ (for general issues concerning Turkish information structure and word order, see Sener 2010). If the direct object is accusative-marked, both acc-dat and dat-acc orders are generally possible for most speakers in an out-of-the-blue context (Issever, p.c.). This is illustrated in (57): (64) a. Can paray-ı adam-a uzattı. Can.NOM money-ACC man-DAT hand-PST.3SG b. ? Can adam-a paray-ı uzat-tı. Can.NOM man-DAT money-ACC hand-PST.3SG 'Can handed the money to the man.' Both sentences can answer the question 'What happened?'. However, Turkish has factors such as specificity which may force a constituent to be placed in a given position for a non-specific reading. This may be the case in (64b) as far as paray-ı ‘money’ is concerned. By contrast, this problem does not arise in (64a), where the DO cannot be interpreted as non-specific. In the absence of more compelling evidence, I will assume that Turkish canonical ditransitive pattern is /DO + IO/, and the two alternates are related by a derivational link, which calls for just one lexical entry for the two alternates. 3.4. Extending the debate to SVOA A parallelism can be established between the behavior of dative alternation and the pattern SVOA, illustrated in (65) in Spanish in contrast with English (66): (65) a. Puse la leche en la nevera. put-PST.1SG the milk in the fridge b. Puse en la nevera la leche. ‘I put the milk in the fridge.’ (66) a. I put the milk in the fridge. b. *I put in the fridge the milk. Only (65a) is natural as a reply for the question ‘What happened?’, showing that the canonical pattern is SVOA, and SVAO is derived in Spanish (Jiménez-Fernández & Vlachos 2019): (67) a. Puse la leche en la nevera. b. # Puse en la nevera la leche. However, (66) shows that in English, the pattern /SVAO/ is simply not part of their grammar (except in cases where the object is too heavy to stay in situ). The heterogeneous nature of verbal alternations Isogloss 2023, 9(4)/6 21 With narrow focus however, we are faced with a different story. In (68) and (69) there is again a preference for one sequence or the other depending on whether the Object or the Adverbial is focused: (68) A: ¿Dónde pusiste la leche? ‘Where did you put the milk?’ B: a. Puse la leche en la nevera. b. # Puse en la nevera la leche. (69) Q: ¿Qué pusiste en la nevera? ‘What did you put in the fridge?’ A: a. # Puse la leche en la nevera. b. Puse en la nevera la leche. The availability of rearrangement or lack thereof in the /SVOA/ pattern may be used to divide languages into two groups, namely languages that exhibit a behavior similar to English and languages which pattern Spanish in licensing this rearrangement: (70) German: Bettina Kaminski, p.c a. Ich stele die Milch in den Kühlschrank. I put-PST.1SG the milk in the fridge b. ??Ich stele in den Kühlschrank die Milch. ‘I put the milk in the fridge.’ (71) French: Lina Elhage-Mensching and Pierre Larrivée, p.c a. Je mets le lait dans le frigo. I put-PST.1SG the milk in the kitchen b. *Je mets dans le frigo le lait. ‘I put the milk in the fridge.’ (72) (Brazilian) Portuguese: Cilene Rodrigues and Roberta Pires de Oliveira, p.c a. Eu coloquei o leite na geladeira. I put-PST.1SG the milk in.the fridge b. Eu coloquei na geladeira o leite. ‘I put the milk in the fridge.’ (73) Greek: Jiménez-Fernández & Vlachos 2019 a. Evala to ghala sto psighio. put-PST.1SG the milk-ACC to-the fridge b. Evala sto psighio to ghala. ‘I put the milk in the fridge.’ Both German and French are of the English-type, whereas Portuguese and Greek are of the Spanish-type. Thus, both orderings are possible, but SVAO is most Isogloss 2023, 9(4)/6 Ángel Jiménez-Fernández 22 natural with narrow focus on the object in the last group, whereas SVOA is preferred if this is an answer to a context question such as Where did you put the milk? 9 To conclude this third section, focusing on Spanish (though the analysis can be extended to other languages), two types of alternations can be held to appear in Spanish, namely verb alternations of the load-type and verb alternations of the givetype. The former entail two separate lexical entries, whereas the latter will require but one single lexical entry. Based on the patterns the alternates use in all-focus sentences, locative alternation will have two syntactic derivations, whereas dative alternation will require just one syntactic derivation and the sequence /OI DO/ will be informationally derived from the canonical /DO IO/. Previous literature has discussed issues concerning the information structure of verbal alternations, as presented in Section 1, especially concerning Dative Alternation (to my knowledge, there is no study about Locative Alternation and discourse). For Brazilian Portuguese, we find references such as Scher (1996), Armelin (2011), Cépeda & Cyrino (2020) or Lacerda (2020); for Spanish, we find Jiménez-Fernández (2009); for English we find Bresnan et al. (2007) or Krifka (2003). The issues tackled in these works are connected with my claim that the use of a given pattern as broad focus is indicative that this pattern is basic (and hence, non-derived) in the relevant language. Most previous literature concentrate on the partition topic+focus and the connection between givenness and syntactic position. They address narrow focus, which I have shown does not tell us anything about a monosemous or polysemous analysis of verbal alternations. They also address broad focus, but they do not associate broad focus with a possible derived or non-derived analysis in syntactic terms of the different types of verbal alternations. My claim is that ordering in Dative and Locative Alternations is crucially dependent on broad focus, and that this discourse strategy discriminates between the two possible analyses of verbal alternations. In my next section, I provide evidence that verbal alternations do not constitute a homogeneous group not only crosslinguistically, but also within one single language. 4. Evidence for the split into two groups: scope freezing Since the seminal paper by May (1977), quantifier scope ambiguities have been used as a basis for structural distinctions. For example, when two quantifiers (universal quantifier ∀+ existential quantifier ∃) are present in a given sentence, ambiguity arises precisely from the presence of those two Qs (Kiss & Pafel 2017). A sentence such as (74) will have two interpretations depending on whether the Q some scopes over every or the other way around (Kiss & Pafel 2017): 9 As for Turkish SVOA, although both the acc-loc and loc-acc orders are possible, the former seems to be better in a broad focus context (Issever, p.c.): (i) a. Aslı kitab-ı masa-da bırak-tı. Aslı.NOM book-ACC table-LOC leave-PST.3SG b.  Aslı masa-da kitab-ı bırak-tı. ‘Aslı left the book on the table.’ The heterogeneous nature of verbal alternations Isogloss 2023, 9(4)/6 23 (74) Some man danced with every woman. ∃>∀/∀>∃ a. There is at least one man (among the men in a given domain) who danced with every woman (in a given domain). b. For every woman (in a given domain), there is at least one (possibly different) man (among the men in a given domain) who danced with her. The interpretation in (74a) is obtained via surface scope. The existential Q some has scope over the universal Q every, which is reflected in syntax by c-command (some c-commands every). However, in (74b), the interpretation is reversed and hence May’s (1977) inverse linking is obtained via Q-raising (inverse scope). In that case, the universal Q every has scope over the existential Q (again reflected by c-command in the syntactic derivation). This situation is replicated in Spanish, as observed in (75): (75) Un gato persigue a cada ratón. ∃>∀/∀>∃ a cat chase-PRS.3SG to each mouse ‘A cat chases every mouse.’ a. A single cat chases every mouse. (∃>∀) b. A different cat chases each mouse. (∀>∃) The presence of two quantifiers in (75) produces an ambiguous interpretation of the sentence. In (75a), we find the surface scope interpretation, in which the existential Q takes scope over the universal Q (∃>∀). However, (75b) shows an inverse scope interpretation, and hence the universal Q takes scope over the existential Q (∀ >∃). 10 Now I turn to the impact of scope and the structural consequences in ditransitives. Based on Aoun & Li (1989), Antonyuk (2020) proposes the Scope Freezing Generalization (SFG): 10 There is some debate about the obligatory distributive nature of the quantifier cada ‘each’. For Leonetti (2007) and Sánchez López (1999), this Q is always distributive (Cada detective resolvió tres crímenes ‘Each detective solved three crimes’). However, NGRAE (2009: 1427-1428) states that the Q cada occurs in two different constructions (closely related to each other), namely distributive and non-distributive. 1) Non-distributive cada: it refers to the total amount of some set of items, be it already introduced or not in the previous discourse (Había sopesado meticulosamente cada palabra de su alocución ‘He weighed every single word of his speech.’). DPs containing non-distributive cada are barred from occurring in subject position (?Cada niño de la clase se había resfriado ‘Each child in the class got a cold.’). If the subject is postverbal, this anomaly disappears (Se había resfriado cada niño de la clase), which may indicate that the interpretation of the Q should be low in the tree. 2) Distributive cada: much more frequent, in this reading the Q is connected to some other expression containing a pronoun or a possessive or some other Q. The combination has a multiplying effect, establishing pairs: (Cada niño entregó su examen ‘Each child handed in their exam.’). I will take it for granted that this quantifier has these two interpretations. Isogloss 2023, 9(4)/6 Ángel Jiménez-Fernández 24 Scope freezing results when one QP raises over another to a ccommanding position within the VP as a result of a single instance of movement. (Antonyuk 2020: 48) Antonyuk (alongside Larson 2014, Cépeda & Cyrino 2020) takes inverse scope as an indication that the ditransitive verb exhibits two alternates which are related by a derivational analysis. In other words, the two verbs show up as two separate lexical entries and there is no structural relation between them. However, if scope freezing is possible, the two verbs in the alternation are related by a derivational analysis and two separate lexical entries are proposed for the two verbs. The diagnostic is described by Cépeda & Cyrino (2020: 105) as follows: When DO contains an existential quantifier (DO∃), IO contains a universal quantifier (IO∀), and the order is DO∃>IO∀, the sentence is scopally ambiguous: it has both a surface and an inverse scope reading. In contrast, when DO contains a universal quantifier (DO∀), IO contains an existential quantifier (IO∃), and the order is IO∃>DO∀, the scope in the sentence is frozen: no inverse scope reading is allowed. Larson (2014) illustrates with the verb teach in English: (76) a. John taught every language to two persons. ∀>∃; ∃>∀ b. John taught two persons every language. ∃>∀; *∀>∃ As observed, in (76a) we obtain two scope-based interpretations. First, ∀>∃ implies that for every language (within a set), there are potentially two (different) people that John taught it to. And secondly, ∃>∀ implies that there are two people that John taught all the languages (maybe in a delimited set) to. However, in (76b), the only interpretation available is the one in which the existential Q takes scope over the universal Q, and hence a reading arises in which each of two people are taught all the languages. As Larson (2014: 84) puts it, “[t]he double object form seems to “freeze” the scope of two quantified object nominals whereas the oblique form does not.” Scope freezing is evidence for the derivational view of Dative Shift in English (also for Russian ditransitives; cf. Antonyuk 2020). 11 The two possible configurations in ditransitives and their interpretatios are summarized in Figure 3: Figure 3. Scope interpretations of double Qs DO∃ + IO∀ >>> Scope ambiguity IO∃ + DO∀ >>> No inverse scope The different scope interpretations are also attested for other verbs such as give in English Dative Shift by Antonyuk & Larson (2016): 11 In relevant current literature on ditransitives, scope is taken as diagnostic of the structural relation existing between the two alternate verbs in a given alternation (Bruening 2010). The heterogeneous nature of verbal alternations Isogloss 2023, 9(4)/6 25 (77) a. The teacher gave a book to every student. (∃>∀, ∀>∃) b. The teacher gave a student every book. (∃>∀, *∀>∃) In (77a), one possible reading is ∃>∀: there is a book that the teacher gave to every student. An alternative interpretation for (77a) is ∀>∃, by which for each student there is a different book that the teacher gave to them. In contrast, in (77b), there is just one reading in which a single student received all the books. This is taken as evidence for a derivational analysis of verbs such as give in English. Cépeda & Cyrino (2020) replicate the same diagnostic for the Dative Alternation in BP: (78) a. A Olga deu [DO um presente] [IO para todos the Olga give-PST.3SG a gift to every os alunos]. (∃>∀, ∀>∃) the students ‘Olga gave a gift to every student.’ b. A Olga deu [IO para um aluno] [DO todos os the Olga give-PST.3SG to a student every the presentes]. (∃>∀, *∀>∃) 12 gifts ‘Olga gave a student every gift.’ (Cépeda & Cyrino 2020: 104, examples 15a-b) In (78a), two interpretations are attested. On the one hand, the ∃>∀ scope implies that that there is a book that the teacher gave to every student. On the other hand, the inverse ∀>∃ scope relation establishes that for each student there is a different book that the teacher gave to them. Conversely, in (78b), only the surface ∃>∀ scope relation is available. Cépeda & Cyrino (2020) take this as evidence that Dative Alternation is BP involves a derivational analysis. Recall that with the all-focus sentence test the pattern in (78a) is the one selected to express broad focus, and hence it is the canonical pattern which may derive into a different pattern for discourse reasons. In other words, scope interpretations and broad focus work hand in hand in supporting a derivational analysis of BP Dative Alternation. I will turn next to Spanish, testing and implementing in this language what Cépeda & Cyrino (2020), Larson (2014) and Antonyuk (2020) have shown in other languages. Spanish Dative Alternation exhibits the same scope interpretations as English and BP: (79) a. Marisa le dio un regalo a cada alumno/a todos Marisa CL-3.SG.DAT give-PST.3SG a gift to each student/to every los alumnos. ∃>∀; ∀>∃ the students 12 An anonymous reviewer points out that the ∀>∃ reading is possible in BP. Variation among speakers is expected regarding the different scope interpretations. However, I will stick to the reading that Cépeda & Cyrino assume in their paper. Isogloss 2023, 9(4)/6 Ángel Jiménez-Fernández 32 why do we also need discourse-based projections? (see discussion in Jiménez- Fernández, Forthcoming). In light of this reductionist idea, I assume that the vP periphery exhibits no designated projections such as Topic Phrase or Focus Phrase (for the original idea in cartography, see Belletti 2001). The relevant features in Spanish enter the derivation in phasal v and are inherited by V. 5.2. Towards a formal analysis In this subsection, I offer a preliminary analysis of the two types of alternations that I have discussed. The analysis for Dative Alternation takes into account the possible information structural role of topic for the DO or the IO as in Cépeda & Cyrino (2020) or Quarezemin (2009), but it does not rely on the presence of a Topic Phrase, but rather on the postulation of a topic feature, much in the spirit of Minimalism. The ingredients of the proposal for Dative Alternation in Spanish are outlined as follows: 1) Ditransitives require the projection of a low ApplP (Pylkkänen 2000; Bleam 2003; Cuervo 2003; Armelin 2011; Fábregas et al. 2016; Jiménez-Fernández 2020). A plausible justification of the presence of this projection is that there seems to be a possessive relation between the DO and the IO. From a sentence such as Juan le dio el libro a María (‘Juan gave the book to Mary’), we all infer that Mary possesses the book. The possession flavor was the original motivation for the inclusion of ApplP, as argued in Pylkkänen (2000). 2) The Appl head is occupied by the dative clitic le (as discussed by Cuervo 2003). 3) To account for the neutral word order, I assume that DO is in spec-ApplP and IO is in the complement of ApplP (contra Cuervo 2003, but following my previous research on datives in Jiménez-Fernández 2020). The main reason why I use this configuration is that it accounts for the canonical word order of Dative Alternation without stipulating any further operation. 4) To explain the all-focus instances, there is no movement except for the clitic to attach the verb. By default, in the absence of discourse features, when the derivation is transferred to the interfaces the sentence is interpreted as broad focus (Bentley & Cruschina 2018). 5) To justify the position of DO in the /DO IO/ sequence, I claim that it has a topic function. As such, the DO undergoes movement to spec-VP after feature inheritance from v to V (Jiménez-Fernández & Spyropoulos 2013; Jiménez- Fernández & Vlachos 2019; Mursell 2021). This does not affect the original word order. 6) To account for the topic interpretation of IO, this argument moves to spec-VP after feature inheritance, rearranging the original word order. The derivation that I propose for the ordering /DO IO/ in (97a) appears in (97b): The heterogeneous nature of verbal alternations Isogloss 2023, 9(4)/6 33 (97) a. Juan le dio el regalo a María. b. vP DP v ' Juan v VP le dio+ø DP V ' el regalo [Top] V ApplP le dio [u-1] DP Appl’ [u-2] el regalo [EF]. [Top] Appl PP le a María [Foc] By Agree V probes the DP el regalo as its Goal. The unvalued discourse feature [u-]1 in V will be valued by the topic feature in the DP el regalo. The Edge Feature (EF) triggers movement of the constituent carrying the discourse feature [Top]. Kratzer & Selkirk (2020) have argued that newness is not associated with a [Foc] feature in English. IF is interpreted after syntax by prosody. However, this is far from common to all languages. There are languages such as Gungbe where IF is marked syntactically but not prosodically (Kratzer & Selkirk 2020). However, in languages such as Spanish, IF is marked only prosodically or by a combination of syntactic position and intonation. The syntactic rearrangement that IF can trigger is indicative that this discourse role involves a -feature, more precisely, an interpretable [Foc] feature. In our derivation, this means that the IO a María is endowed with the feature [Foc] (Aboh 2010), which values the [u-]2 under V. No movement is required since the EF under V has already been satisfied. Now let’s turn to the pattern /IO DO/ and the discourse features present in the low periphery of the construction. This is illustrated in (98a), for which I put forth the derivation in (98b): 14 14 There are differences in terms of word order and the definite or indefinite nature of the direct object in English. Krifka (2008) identifies some preference of a particular alternation over the other one: (i) a. Bill showed the boy a girl. b. *Bill showed a boy the girl. c. Bill showed the girl to a boy. The selection of a pattern depends on the occurrence of the indefinite DP in last position, which favors a focus reading. Thus, (ib) violates the Information Flow Principle in that it Isogloss 2023, 9(4)/6 Ángel Jiménez-Fernández 34 (98) a. Juan le dio a María el regalo. b. vP DP v ' Juan v VP le dio+ø PP V ' a María [Top] V ApplP le dio [u-1] DP Appl’ [u-2] el regalo [EF] [Foc] Appl PP le a María [Top] The unvalued discourse feature [u-1] in V will be valued by the topic feature in the PP a María. The EF triggers movement of the Topic-marked element to spec- VP. However, [u-2] is valued by Agree with the focused DO el regalo. Thus, we obtain the ordering /IO DO/. Now we move to all-focus ordering in Dative Alternation. For broad focus, we have seen that the natural ordering is /DO IO/, as illustrated in (99a). The syntactic analysis I proposed for this all-focus sentence is provided in (99b): (99) a. Juan le dio el regalo a María. (all-focus) b. vP DP v ' Juan v VP le dio+ø V ApplP le dio DP Appl’ el regalo Appl PP le a María Here, for economy reasons, no IS-related features are necessary, and V is not endowed with an EF. Hence, no movement is involved. The all-focus interpretation is obtained after syntax (Fábregas et al. 2015), at the interfaces. When the interpretative presents new information following background information (Chafe 1976). These definiteness effects are evidence for an analysis based on the /topic + focus/ partition. The heterogeneous nature of verbal alternations Isogloss 2023, 9(4)/6 35 levels realize that there are no discourse features to be interpreted, the broad focus interpretation obtains (Bentley & Cruschina 2018). The next step is the computation of Locative Alternation in Spanish. The ingredients of my proposal for Spanish Locative Alternation consider that two separate computations are necessary. These ingredients are listed below: 1) No ApplP proper is required; rather the two postverbal arguments are generated inside a small clause. (Kayne 1984; Pesetsky 1995; Harley 1996; 2002; Cuervo 2003; Beck & Johnson 2004; Jung & Miyagawa 2004; Harley & Miyagawa 2017; McIntyre 2006; Pylkkänen 2008; Schäfer 2008; Harley & Jung 2015; Bondarenko 2018, among others). The reason why there is no ApplP is that there is no possessive meaning which may connect the two postverbal arguments. For example, from María untó mermelada en la tostada (‘Mary spread jam on toast’) we infer a locative relation where the location is the slices of bread, but not a real possessive relation. 2) The alternation may be reduced to the syntactic position (specifier or complement) that each participant occupies in the SC (FP). Positing this SC is supported by the locative relation that is established between the two participants (the jam is on toast), based on Kayne’s (1984) line of argumentation according to which location is a SC predicate. 3) The two alternates have a different neutral ordering for all-focus interpretation. 4) As stated earlier, for the broad focus interpretation no discourse features are involved in the syntactic computation. All-focus ordering for each alternate is illustrated in (100). For (100a) I put forth the analysis in (101), whereas for (100b) I provide the analysis in (102): (100) a. María untó mermelada en la tostada. b. María untó la tostada con mermelada. (101) vP DP v ' María v VP untó+ø V FP untó DP F’ mermelada F PP en la tostada (102) vP DP v ' María v VP untó+ø V FP Isogloss 2023, 9(4)/6 Ángel Jiménez-Fernández 36 untó DP F’ la tostada F PP con mermelada In both cases, no IS-related features are involved since the whole sentence is interpreted as broad focus. V has no EF; therefore, no movement takes place. As with Dative Alternation, the all-focus interpretation in Locative Alternation is obtained after syntax (Fábregas et al. 2015). In case of a different discourse interpretation (for example, if a /topic + focus/ partition is involved in the postverbal sequence), the relevant discourse features will be inherited from v to V and the derivation would be exactly as the ones presented for Dative Alternation above. One final note is in order regarding the topic interpretation of the postverbal participant in both Dative and Locative Alternations. I have mentioned that Quarezemin (2009), Cépeda & Cyrino (2020), Jiménez-Fernández (2009) and Lacerda (2020) propose the projection of a TopP in the low periphery, assuming cartography. The analysis I propose is based on discourse features with no stipulation of dedicated projections. For ditransitives, where the postverbal sequence reflect a /topic + focus/ partition, both approaches are theoretically sound. However, in the minimalist spirit that I am assuming, my analysis is preferred in that no redundancy or superfluous projections are available. 6. Conclusions In this paper I have discussed the nature of two verbal alternations, namely Dative Alternation and Locative Alternation. Traditionally, two different approaches have emerged. Alternate verbs are derivationally or non-derivationally related. However, by analyzing data from different languages (English, Turkish, Portuguese and especially Spanish), we may notice that the nature of alternations is not homogeneous both cross-linguistically and language-internally. In Spanish, we find at least two types of ditransitives: 1) Alternations which involve a derivational relation between the two alternates, and 2) Alternations which require no relation between the two alternates. Information Structure plays a crucial role in understanding some types of alternations (Dative Shift): a 𝛅-feature may work in conjunction with an EF, triggering topic movement. Broad focus has been used as a test to reveal whether the ordering of objects is canonical or derived. This brings about consequences for linguistic theory in that all-focus sentences have been observed to be a reliable diagnostic for tearing apart the two groups. Scope-freezing has proved to be confirmation of the different behavior of the two groups. Acknowledgments The research here has been partially funded by research project PID2022-137233NB- I00 of Spain’s Ministry of Science, Innovation, and Universities (MICINN). I am most thankful to colleagues, especially Cilene Rodrigues and Roberta Pires de Oliveira for their help with data from Brazilian Portuguese and for their kind invitation to The heterogeneous nature of verbal alternations Isogloss 2023, 9(4)/6 37 participate in this special issue, and Selçuk İșsever for his comments on Turkish data. I also thank Aimée Lewis for revising the English. List of abbreviations used in the article ACC accusative CL clitic DAT dative DO direct object IO indirect object NOM nominative PL plural PRES present PST past SG singular References Aboh, Enoch. 2010. Information structure begins with the numeration. Iberia: An international journal of theoretical linguistics 2: 12–2. Agenjo, Dolores. 2019. Los verbos con alternancia locativa transitiva en español: componentes semánticos y estructuras argumentales. PhD. thesis, Universidad Carlos III, Madrid. Anderson, Stephen. 1971. On the role of Deep Structure in semantic interpretation. Foundations of language 6: 197-219. Antonyuk, Svitlana. 2020. The puzzle of Russian diatranstivies. In Anna Pineda & Jaume Mateu (eds.), Dative constructions in Romance and beyond, 43–74. Berlin: Language Science Press. Antonyuk, Svitlana & Richard Larson. 2016. Scope freezing in PP dative constructions? Presentation given at Formal Description of Slavic Languages 12 (FDSL), Humboldt University. Antonyuk, Svitlana & Roksolana Mykhaylyk. 2022. Scope freezing and object shift in Ukrainian: Does superiority matter? Remark. Syntax 25(1): 122-146. https://doi.org/10.1111/synt.12229 Aoun, Joseph & Yen-Hui A. Li. 1989. Scope and constituency. Linguistic inquiry, 20(2): 141-172. Isogloss 2023, 9(4)/6 Ángel Jiménez-Fernández 38 Armelin, Paula. 2011. Sentenças bitransitivas do portugués de Brasil revisitadas à luz da teoría de núcleos funcionais aplicativos. MA dissertation, University of Sao Paulo. Baker, Mark C. 1997. Thematic roles and syntactic structure. In Liliane Haegeman, ed., The elements of grammar, 73-137. Dordrecht: Kluwer. Belletti, Adriana. 2001. Inversion as focalization. In Aafke Hulk & Jean-Yves Pollock eds.), Subject inversion in Romance and the theory of universal grammar, 60–90. New York, NY: Oxford University Press. Bleam, Tonia. 2003. Properties of the double object construction in Spanish. In Rafael Núñez-Cedeno, Luis López & Richard Cameron (eds.), A Romance perspective on language knowledge and use, 233–252. Amsterdam/Philadelphia: John Benjamins. Beck, Sigrid & Kyle Johnson. 2004. Double objects again. Linguistic inquiry 35(1): 97–124. https://doi.org/10.1162/002438904322793356 Bentley, Delia & Silvio Cruschina. 2018. The silent argument of broad focus: Typology and predictions. Glossa: a journal of general linguistics 3(1): 118. 1–37. https://doi.org/10.5334/gjgl.677 Bondarenko, Tatiana. 2018. Russian datives again: On the (im)possibility of the small clause analysis. In Denisa Lenertová, Roland Meyer, Radek Šimík & Luka Szucsich (eds.), Advances in formal Slavic linguistics 2016, 25–51. Berlin: Language Science Press. https://doi.org/10.5281/zenodo.2545511 Bruening, Benjamin. 2010. Double object constructions disguised as prepositional datives. Linguistic inquiry 41(2): 287–305. https://doi.org/10.1162/ling.2010.41.2.287 Cépeda, Paola & Sonia Cyrino. 2020. Putting objects in order: Asymmetrical relations in Spanish and Portuguese ditransitives. In Anna Pineda & Jaume Mateu (eds.), Dative constructions in Romance and beyond, 97–116. Berlin: Language Science Press. Chafe, Wallace. 1976. Givenness, contrastiveness, definiteness, subjects, topics, and point of view. In Charles Li (ed.), Subject and topic, 25–55. New York, NY: Academic Press. Chafe, Wallace. 1987. Cognitive constraints on Information Flow. In Russell Tomlin (ed.), Coherence and grounding in discourse, 21–51. Amsterdam: Benjamins. Chomsky, Noam. 2001. Derivation by phase. In Michael Kenstowicz (ed.), Ken Hale: A life in language, 1–52. Cambridge, MA: MIT Press. Chomsky, Noam. 2007. Approaching UG from below. In Uli Sauerland & Hans Martin Gärtner (eds.), Interfaces + Recursion = Language? Chomsky's Minimalism and the View from Syntax-Semantics, 1–30. Berlin: Mouton de Gruyter. The heterogeneous nature of verbal alternations Isogloss 2023, 9(4)/6 39 Chomsky, Noam. 2008. On Phases. In Robert Freidin, Carlos Otero & María Luisa Zubizarreta (eds.), Foundational Issues in Linguistic Theory, 133–166. Cambridge, MA: The MIT Press. Cifuentes Honrubia, José Luis. 2008. Removal verbs and locative alternations in Spanish. ELUA 22: 37-64. http://dx.doi.org/10.14198/ELUA2008.22.03 Contreras, Heles. 1983. El orden de palabras en español. Madrid: Cátedra. Cuervo, María Cristina. 2003. Datives at large. Ph.D. thesis, Cambridge, Massachusetts Institute of Technology. https://dspace.mit.edu/handle/1721.1/7991. Demonte, Violeta. 1995. Dative alternation in Spanish. Probus 7(1): 5–30. https://doi.org/10.1515/prbs.1995.7.1.5. Enç, Mrvet. 1991. The semantics of specificity. Linguistic inquiry 22: 1–25. Erteschik-Shir, Nomi. 2007. Information structure. Oxford: Oxford University Press. Fábregas, Antonio, Ángel L. Jiménez-Fernández and Mercedes Tubino Blanco. 2017. 'What's up' with datives?. In Ruth Lopes, Juanito Avelar & Sonia Cyrino (eds.), Romance languages and linguistic theory 12. Selected papers from the 45th linguistic symposium on Romance languages (LSRL), Campinas, Brazil, 29–48. Amsterdam: John Benjamins. https://doi.org/10.1075/rllt.12.03fab Fedriani, Chiara & Maria Napoli. 2023. The missing dative alternation in Romance: Explaining stability and change in the argument structure of Latin ditransitives. Transactions of the Philological Society: 1–32. https://doi.org/10.1111/1467- 968X.12255 Fernández-Soriano, Olga. 1999. Two types of impersonal sentences in Spanish: Locative and dative subjects. Syntax 2(2): 101–140. https://doi.org/10.1111/1467- 9612.00017 Fillmore, Charles J. 1970. The grammar of hitting and breaking. In Roderick A. Jacobs & Peter S Rosenbaum (eds.), Readings in English transformational grammar, 120- 133. Waltham, MA: Ginn. Georgala, Effi. 2011. Scrambling verb-final languages and the underlying order of objects in ditransitive constructions. University of Pennsylvania Working Papers in Linguistics 17(1), 105–114. https://repository.upenn.edu/pwpl/vol17/iss1/13 Goldberg, Adelle. E. 1995. Constructions: A construction grammar approach to argument structure. Chicago: University of Chicago Press. Gussenhoven, Carlos. 2007. Types of focus in English. In Chungmin Lee & Matthew Gordon, Topic and focus: Cross-linguistic perspectives on meaning and intonation, 83-100. New York: Springer. https://doi.org/10.1007/978-1-4020-4796-1_5 Isogloss 2023, 9(4)/6 Ángel Jiménez-Fernández 40 Harley, Heidi 1995. Subjects, events, and licensing. Ph.D. thesis, Massachusetts Institute of Technology. Harley, Heidi & Hyun Kyoung Jung. 2015. In support of the PHAVE approach to the double object construction. Linguistic Inquiry 46(4): 703–730. https://doi.org/10.1162/LING_a_00198 Harley, Heidi & Shigeru Miyagawa. 2017. Syntax of ditransitives. In Oxford research encyclopedia of linguistics. Oxford University Press. https://doi.org/10.1093/acrefore/9780199384655.013.186 Issever, Selçuk. 2003. Information structure in Turkish. Lingua 113: 1025–1053. https://doi.org/10.1016/S0024-3841(03)00012-3 Jackendoff, Ray. 1990. Semantic structures. Current studies in linguistics 18. Cambridge, MA: MIT Press. Jiménez-Fernández, Ángel L. 2009. The low periphery of double object constructions in English and Spanish. Philologia hispalensis 23: 179–200. Jiménez-Fernández, Ángel L. 2010. Discourse–agreement features, phasal C and the edge: a minimalist approach. Diacrítica – Language sciences series 24(1): 25–49. Jiménez-Fernández, Ángel L. 2020. Syntax-information structure interactions in the sentential, verbal and nominal peripheries. Newcastle: Cambridge Scholars Publishing. Jiménez-Fernández, Ángel L. 2023. English/Spanish relatives and their relative information structure: A view from language contact in Puerto Rico. Catalan journal of linguistics 22: 29–70. https://doi.org/10.5565/rev/catjl.339 Jiménez-Fernández, Ángel L. Forthcoming. Discourse in Minimalism. In Evelina Leivada & Kleanthes Grohmann (eds.), The Cambridge handbook of minimalism and its applications. Cambridge: Cambridge University Press. Jiménez-Fernández, Ángel L. & Vassilios Spyropoulos. 2013. Feature inheritance, vP phases and the information structure of small clauses. Studia linguistica 67 (2): 185– 224. https://doi.org/10.1111/stul.12013 Jiménez-Fernández, Ángel L. & Bożena Rozwadowska. 2016. The information structure of dative Experiencer psych verbs. In Bożena Cetnarowska, Marcin Kuczok, & Marcin Zabawa, (eds.) Various dimensions of contrastive studies, 100–121. Katowice, Poland: Wydawnictwo Uniwersytetu Śląskiego: Jiménez-Fernández, Ángel. L. & Bożena Rozwadowska. 2017. On subject properties of datives in psych predicates. A comparative approach. Acta linguistica hungarica 64(2): 1-24. https://doi.org/10.1556/2062.2017.64.2.4 The heterogeneous nature of verbal alternations Isogloss 2023, 9(4)/6 41 Jiménez-Fernández, Ángel L. & Christos Vlachos. 2019. Discourse features, vP- periphery and rearrangement: Evidence from Spanish, Greek and English. Talk delivered at the 14th International Conference of Greek Linguistics. University of Patras, Greece. Jung, Yeun-Jin & Shigeru Miyagawa. 2004. Decomposing ditransitive verbs. In Proceedings of the Seoul international conference on generative grammar, 101–120. Seoul: Hansung University. https://static1.squarespace.com/static/5728dfa81bbee0da7943b811/t/5799a68acd0f68 5197925ae8/1469687435655/Korean+DO+SICGG.pdf Kato, Mary A & Francisco Ordóñez. 2019. Topic-subjects in Brazilian Portuguese and clitic left dislocation in Dominican Spanish, clitics and null subjects. Syntax 22: 229- 247. https://doi.org/10.1111/synt.12180 Kayne, Richard. 1984. Connectedness and binary branching. Dordrecht: Foris. https://doi.org/10.1515/9783111682228 Kearns, Kate. 2011. Semantics, 2nd edition. Basingstoke, United Kingdom: Palgrave Macmillan. Kornfilt, Jaklin. 2003. Scrambling, subscrambling and case in Turkish. In Word order and scrambling, In Simin Karimi (ed.), 125–155. Malden, MA: Blackwell. https://doi.org/10.1002/9780470758403.ch6 Kratzer, Angelika & Lisa Selkirk. 2020. Deconstructing information structure. Glossa: a journal of general linguistics 5(1): 113. https://doi.org/10.5334/gjgl.968 Krifka, Manfred. 2003. Semantic and pragmatic conditions for the dative alternation. Paper presented at KASELL International Conference on English Language and Linguistics. Seoul, June 25-26. https://amor.cms.huberlin.de/~h2816i3x/Publications/DativeAlternationKorea.pdf Krifka, Manfred. 2008. Basic notions of information structure. Acta Linguistica Hungarica 55 (3–4): 243–276. https://doi.org/10.1556/aling.55.2008.3-4.2 Lacerda, Renato. 2017. Information structure in child English: Contrastive topicalization and the dative alternation. In Maria LaMendola & Jennifer Scott (ed.), Proceedings of the 41st annual Boston University conference on language development, 387–400. Somerville, MA: Cascadilla Press. Lacerda, Renato. 2020. Middle-field Syntax and information structure in Brazilian Portuguese. Ph.D. Thesis, University of Connecticut. Larson, Richard K. 1988. On the double object construction. Linguistic Inquiry 19(3): 335–391.