On the Decline of Pleonastic "that" in Late Middle English and Early Modern English

1 ‘Althogh that summe thynke the harte to be the sprynge of blood’ On the Decline of Pleonastic that in Late Middle English and Early Modern English1 Javier Calle Martín University of Málaga Abstract The origin of pleonastic that can be traced back to Old English where it could appear in syntactic constructions consisting of a preposition + demonstrative pronoun (i.e. for þy þat, for þæm þe) or a subordinator (i.e. oþ þat). Its diffusion with other subordinators is considered an early Middle English development as a result of the standardization of this item as the general subordinator in the period, which motivated its use as a pleonastic word in combination with all kinds of conjunctions (i.e. now that, gif that, when that, etc.) and prepositions (i.e. before that, save that, in that) (Fischer 1992: 295). Its use considerably increased in late Middle English, reaching its climax in the 16th century. In the 17th century, however, its use rapidly declined, to such an extent that it was virtually obliterated at the end of that same century (Rissanen 1999: 303-304). The list of subordinating elements includes relativizers (i.e. this that), adverbial relatives (i.e. there that) and a number of subordinators (i.e. after, as, because, before, beside, for, if, since, sith, though, until, when, while, etc. The decline of pleonastic that, however, is not a simultaneous process with all these subordinators insofar as the subordinator for seems to have retained this pleonastic item longer than other conjunctive words. In the light of this, the present paper has been conceived with the following objectives: a) to analyse the use and distribution of pleonastic that in a corpus of early English medical writing (in the period 1375–1700); b) to classify the construction in terms of the two different varieties of medical texts, i.e. treatises and recipes; and c) to assess the decline of the construction with the different conjunctive words. The data used as sources of evidence come from The Corpus of Early English Medical Writing, i.e. Middle English Medical Texts (MEMT for the period 1375–1500) and Early Modern English Medical Texts (EMEMT for the period 1500– 1700). 1. Introduction Derived from the Greek term πλεονασµος ‘superabundance, excess’, a pleonasm is used in Linguistics to denote a particular type of redundant linguistic expression, either as a fault of style, or as a figure purposely used for special force or clearness (OED s.v. pleonasm, n.). In many cases, English in particular, these pleonasms function as a rhetorical repetition for a specific linguistic effect, where the redundancy can be dropped with no loss of meaning. Pleonasms can be syntactic (i.e. I know that she will be here soon, the reason is because) or strictly semantic (i.e. a free gift, a true fact, aches and pains, null and void, tuna 1 The present research has been funded by the Ministry of Economy and Competitiveness (grant number FFI2014-57963-P), and by the Autonomous Government of Andalusia (grant number and P11-HUM7597). I am particularly grateful to Prof Matti Rissanen (University of Helsinki, Finland) for his information about the Old Norse influence on the phenomenon and for his comments on previous drafts of this paper. 2 fish, at 12 midnight, etc.).2 The phenomenon has been differently labelled in the literature, from hyperdetermination (Malkiel 1957: 79) to hypercharacterization (Lehman 2005: 119) and overcharacterization (Booij 2007: 273), even though referring to different types of redundant constructions. To my knowledge, the most systematic treatment of the topic is provided by Lehman who classifies pleonasm at different levels of linguistic analysis, identifying hypercharacterization as “pleonasm at the level of grammar” (2005: 119). In the light of his categorization, the phenomenon can be safely classified into tautology (i.e. one constituent is synonymous with the other); stylistic pleonasm (i.e. pragmatically motivated) and hypercharacterization (i.e. grammatically motivated), regardless of whether it responds to semantic and morpho-syntactic motivations.3 The use of pleonastic that in the history of English can be safely ascribed to Lehman’s category of hypercharacterization according to which the general subordinator that becomes spuriously added to all kinds of conjunctions and prepositions (Fischer 1992: 295). On historical grounds, its origin can be traced back to Old English where it could appear in syntactic constructions consisting of a preposition + demonstrative pronoun (i.e. for þy þat, for þæm þe) or a subordinator (i.e. oþ þat). These combinations simplified in Middle English but that was kept after conjunctions, appended even to conjunctions with which it did not appear in Old English (i.e. gif that, when that, etc.), prepositions also included (i.e. before that, save that, in that). Its use considerably increased in late Middle English, reaching its climax in the 16th century, with is progressive obliteration during the course of the 17th century (Rissanen 1999: 303-304), as shown in examples (1-2). (1) and y beleue ȝif that he wyl take hede to alle […] (lanfranc_chirurgia_magna_2.txt). (2) The man that is borne while that the Moon enters in this sign (1539_Moulton_MyrrourOrGlasseOfHelth_Astrology_NORM.txt). Different interpretations account for the development of this phenomenon in Middle English, one considering it as a native and the other as a non-native phenomenon. On the one hand, the rise of pleonastic that is viewed as the result of the standardization of that as the general subordinator in Middle English, the list including nominal, adjectival and particular types of adverbial clauses. This heterogeneous nature of the subordinator that plausibly contributed to its extension to other types of subordinate 2 The emphatic element of pleonasms turns them into a recurrent device in many languages. Spanish, among others, count on instances such as subir para arriba, bajar para abajo, el día martes, el mes de septiembre, hoy en día, hoy por hoy, etc. In French, the most typical example of pleonasm is the use of the word pas with the negative adverb ne for strictly emphatic purposes (Wigdorsky 2004: 171-178; also Iglesias-Casal 1989: 118-122). 3 The term overcharacterization is elsewhere considered to be more lexically-oriented to refer to the addition of a superfluous suffix in the coining of a new word, the terms hyperdetermination and hypercharacterization have a morpho-syntactic bias (Gardani 2015: 538). 3 clauses where it was not strictly necessary, becoming a form with currency in the latter part of Middle English (Kivimaa 1966: 248; Fischer 1992: 295). Its development, on the other hand, has also been conceived from the point of view of language contact situation as a result of a likely Scandinavian influence (Rissanen 1999: 303).4 Regardless of the ultimate origin of the construction, it is also a fact that the particular requirements of metre eventually paved the way to its standardization. In this vein, it is considered in many cases as a line filler used to achieve metrical regularity, a fact which explains why Chaucer uses the phenomenon twice as common in verse than in prose (1968: 48-49; 7071). Pleonastic and non-pleonastic that are not, however, clearly distinguished in Chaucer, eventually producing the unconscious incorporation of this spurious element in his prose (Brinton 1996: 352; also Potter, Jefferson and Stokes 2007: 135). In our opinion, there is still an important gap in the field, especially from a historical perspective as the phenomenon has been mostly discussed from a synchronic perspective in Late Middle English, Chaucer’s verse in particular. The present paper then analyses the origin, development and decline of pleonastic that from a diachronic point of view in the light of The Corpus of Early English Medical Writing in the historical period 1375-1700, considering that medical writing is freer from the artificiality of verse and may offer some fresh details about the ups and downs of this pleonastic form. By the year 1475, the first phase of the vernacularization of scientific and medical writing was largely complete, to such an extent that the use of the vernacular became even more common than Latin for the rendering of scientific material (Voigts 1986: 316; 1996: 816; Pahta & Taavitsainen 2004: 12). In the absence of a national standard for these purposes, the emergence of the scientific register, shaped under the shelter of the Greco-Roman models, may shed some light on the development of these on-going changes in English. In the light of this, the present paper has been conceived with the following objectives: a) to analyse the use and distribution of pleonastic that in a corpus of early English medical writing (in the period 1375–1700); b) to classify the construction in terms of the different types of medical texts; and c) to assess the decline of the construction with the different conjunctive words. 2. Methodology The data used as source of evidence come from the two sections of the Corpus of Early English Medical Writing: Middle English Medical Texts (henceforth MEMT), for the historical period 1375-1500 and Early Modern English Medical Texts (henceforth EMEMT), for the historical period 1500-1700. These corpora have been chosen both on quantitative and qualitative grounds. In quantitative terms, they amount to more than 1.8 4 Old Norse at could be used pleonastically in the same way as the earlier English that. This use is mentioned in A Dictionary of Old Norse Prose giving an example of the phrase ef at introducing a subordinate clause. This pleonastic use of at is also included in the glossary of Gordon’s Introduction to Old Norse (1957), providing references to text passages. 4 million words, which is a sizeable input for the analysis of this type of constructions. From a qualitative perspective, on the other hand, the Corpus of Early English Medical Writing is organised into three main branches: specialised treatises, surgical treatises and recipe collections, the latter of a more colloquial nature (Pahta and Taavitsainen 2004: 7). Corpus Specialized texts Surgical texts Remedies Total MEMT 88,349 137,794 219,395 445,538 EMEMT 762,667 298,352 339,068 1,400,087 EMEMT1 1500–1549 59,602 21,910 46,814 128,326 EMEMT2 1550–1599 162,313 102,919 92,405 357,637 EMEMT3 1600–1659 228,135 50,771 71,047 349,953 EMEMT4 1650–1700 312,617 122,752 128,802 564,171 Table 1. Word count for MEMT and EMEMT MEMT contains more than half a million words based on both edited medical texts and early printed books from 1375 to 1500. The bulk of these treatises are translations from Latin, dealing with a wide variety of topics, such as ophthalmology, gynaecology, urinoscopy, phlebotomy, epilepsy, syphilis and the plague (Méndez-Naya and Pahta 2010: 193). EMEMT, in turn, is a two-million-word corpus of medical writing for the period 1500-1700 covering “the full range of printed medical writing in the early Modern period, with its rich diversity” (Taavitsainen and Tyrkkö 2010a: 57). Following the late Middle English component, EMEMT is divided into “theoretical treatises”, “surgical and anatomical treatises” and “remedies” (Taavitsainen and Tyrkkö 2010b: 65-66; Pahta and Ratia 2010: 73-74; Marttila 2010: 102-103; Tyrkkö 2010: 119-120). For comparison, the EMEMT material has been classified into four sub-periods of fifty years each based on the year of printing indicated in the sources. Table 1 below reproduces the word count for the source data.5 MEMT EMEMT Specialized texts General Treatises or textbooks Treatises on specific topics Surgical texts Surgical and anatomical treatises Remedies and materia medica Recipe collections and materia medica Regimens and health guides Philosophical Transactions 5 General treatises or textbooks provide a systematic account of the whole field of medicine; treatises on specific topics are texts on particular diseases and concrete methods of diagnosis or treatment, on specific therapeutic substances, on midwifery and children’s diseases and on the plague; recipe collections and materia medica contain both remedy books and formalized pharmacopoeias; regimens and health guides comprise texts on preventive medicine; surgical and anatomical treatises constitute treatises on surgery and anatomy; and Philosophical Transactions consist exclusively of medical texts published as letters or articles in the Philosophical Transactions of the Royal Society (Taavitsainen et al. 2011: 22-25). 5 Table 2. Genre categorization of MEMT and EMEMT AntConc (version 3.4.3) has been used for the automatic retrieval of the instances. The process, however, was not straightforward in view of the particulars of the corpora. As the tagged version of MEMT is not hitherto available, the present research has been based on the plain text version of the corpus. For convenience, however, the complete concordances for that was generated and sorted according to the preceding context. This facilitated the disambiguation of the instances in such as way the instances without a preceding conjunction could then be easily discarded. As for EMEMT, on the other hand, the tagged version of the corpus has been used, which has allowed us to retrieve all the instances of pleonastic that automatically, just by prompting the particular conjunction involved. The present study is restricted to the analysis of pleonastic that in relative clauses (i.e. which and who) and the following subordinators (i.e. afore, after, as, because, before, beside, for, if, since, sith, though, until, what, when, where, while, why, etc.). The process, however, was not straightforward inasmuch there were as a number of instances where that did not appear in pleonastic use, either as a deictic demonstrative or as a conjunction introducing a that-clause, as shown in examples (3-4). The disambiguation process thus trimmed down the corpus instances to a total of 1,647 instances, 602 in MEMT and 1,045 in EMEMT. (3) […] be it so known that this or that medicine, hath virtue to fill an hollow (1586_Galen_MethodusMedendi_NORM.txt). (4) Neither is it to be doubted but that these humours doe move (1640_Brugis_MarrowOfPhysicke_NORM.txt). 3. Analysis 3.1. Quantitative analysis The present section deals with the diachronic development of the phenomenon in the historical period 1375-1700. Table 3 thus reproduces the number of instances where the figures are offered both in raw and normalized figures to a text of 10,000 words for comparison. The results tentatively confirm that the use of pleonastic that was well diffused in Late Middle English with a distribution of 135.12 instances in MEMT. Notwithstanding this, the construction is observed to decline slightly throughout the early modern English period, with 80.65 and 91.39 instances in-between 1500-1549 and 1550-1599, respectively. The 17th century witnesses the progressive obliteration of the pleonastic use of that to such an extent that the phenomenon amounts to just 30.86 instances at the turn of that same century, one third of the instances than in the previous century. Raw n.f. 1375-1500 602 135.12 1500-1549 165 80.65 6 1550-1599 388 91.39 1600-1649 255 63.48 1650-1700 237 30.86 Total 1,647 Table 3. Distribution of pleonastic that across time (raw and n.f.) This picture obtained from early English medical writing does not strictly coincide with the distribution of pleonastic that in other typology of texts. In his analysis of The Helsinki Corpus of English Texts, Rissanen describes the phenomenon as a relatively common device in the 16th century with a decreasing popularity in the 17th century, its use stretching until the beginning of the following century (Rissanen 1999: 303). Medical writing, in turn, confirms the actual rise and spread of the phenomenon as a Late Middle English development in view of its occurrence in MEMT, with 135.12 occurrences in contrast with 91.39 in the second half of the 16th century. This Late Middle English preference is plausibly justified in the light of the recurrent use of pleonastic that in combination with the subordinators for and if, especially if compared with other conjunctive words. The Corpus of Early English Medical Writing also allows us to classify the construction in terms of the different text-types, i.e. specialized texts, surgical texts and remedies (see Table 2 above). In addition, a new genre emerged towards the end of the EModE period, the Philosophical Transactions, which embodies the state of scientific writing after the transition from a scholastic to an empirical thought-style (Taavitsainen and Pahta 2011; see also Gotti 2001; Taavitsainen 2002)6. In other words, these genres represent different registers within the field of medical writing. Thus, while Theoretical Treatises are considered the most academic register, remedies portray the language used by laypeople, as they were collection of recipes that families stored for their use at home. Surgical Treatises, in turn, would fall in-between the above-mentioned classes. Figure 1 reproduces the distribution of this pleonastic item across the different text types, where the figures have been normalized for the sake of comparison. In MEMT, on the one hand, the construction is found to diffuse more substantially in surgical treatises (with 260.53 occurrences) followed by theoretical treatises (with 235.43 occurrences). More important, however, is the constrained distribution of the phenomenon in remedies where the use of pleonastic that plunges to just 15.95 occurrences every 10,000 words. These figures tentatively lead us to formulate on the likely impulse of this pleonasm in the late Middle English period. Whilst the use of that in these contexts becomes a recurrent feature in both surgical and theoretical treatises, its occurrence is erratic 6 Note that the category Philosophical Transactions is only present in EMEMT, as it began to be produced towards the end of the 17th century. Although no ME counterpart is available, we consider that the features of this genre are important insofar as it would ultimately define the characteristics of modern scientific English (Taavitsainen 2011). 7 remedies. This fact may shed some new light on the development of the construction in the latter part of Middle English. Remedy books can be traced back to the Old English period, and consist of treatments for ailments written by non-practitioners based on “adaptation and accretion”, ultimately devised for the use of layman and academic phisicians (Voigts and McVaugh 1984: 21). Theoretical and surgical treatises, in turn, were new in the Middle English period and belonged to the learned tradition, being mostly translations of learned Latin medicine with an academic origin designed for physicians of the highest class, surgeons and barber surgeons.7 Their different traditions of writing and intended readership may, to a certain extent, justify the heterogeneous condition of pleonastic that in these texts. My argument then complies with Kivimaa’s point as to the artificial nature of this pleonasm, becoming considerably more frequent in Chaucer’s verse rather than his prose (Kivimaa 1968: 48-49; 70-71). Medical writing also corroborates this same state of affairs in the sense that the construction becomes more frequent in translated material composed under the shelter of learned circles, this pleonasm being a kind of hypercorrected form coming from the Latin conjunction quid, or its French counterpart que, as shown in examples (5 and 6). Remedies, on the contrary, are free from the above-mentioned artificiality of learned compositions, and consequently they do not make enough room for this pleonastic item in the period 1350-1500 when the construction was still incipient, plausibly with a constrained role in the spoken domain. (5a) [...] be flebotomye is not mynystrid materie to þe placi putride, or for þat be flebotomye þe veynes are voide [and] by cause of voidenesse ar drawe humourus wiþ outforþ contened (Voigts and McVaugh 1984: 45, 158-160). (5b) [...] per flebotomiam eductis non administratur materia loco putrido, vel quia per flebotomiam vene evacuantur et gratia evacuationis attrahunt humores extra vasa existentes (Voigts and McVaugh 1984: 44, 158-160). (6a) The best of all to purge downward is the marmellade of quinces of Lion, whereunto hath been added a litle myrobolans citrins brought to poulder. After that the humour hath been purged, to comfort the stomack and bind the vpper mouth (Treatise of the Diseases of Women, GUL 303, 17h-century). (6b) Le meilleur de tous pour purger par bas, est le cotignac de Lyon, aquel on aura adioutté quelque peu de myrobalans citrins puluerisez. Apres que l’bumeur fera purgé, pour conforter l’estomach, & astreindie de plus en plus son orifice superieur (Jean Liébault Trois Libre Appartenans aux Infirmitez et Maladies des Femmes, 1585, p. 21, chapter X). 7 Taavitsainen refers to the problematic position of surgical treatises in Voigts and McVaugh’s taxonomy (1984). Even though they also to the academic tradition, these texts mostly originate in learned circles, they have a place among practical sciences as well (Taavitsainen and Pahta 1997: 74). 8 Fig 1. Distribution of pleonastic that across text types (n.f.) EMEMT, on the other hand, shows a different state of affairs. Even though surgical and theoretical treatises are still found to have the highest number of instances with a total 65.69 and 59.38 occurrences, remedies show a substantial increase if compared with the Late Middle English distribution amounting to 55.74 occurrences, followed by philosophical transactions with 46.31 occurrences. The 16th century witnesses the actual rise of the phenomenon, reaching its climax towards the end of that same century (Rissanen 1999: 303), an evidence corroborated by the rise of the construction in recipes, which is considered a popular text-type if compared with theoretical and surgical treatises. In itself, it may be safely considered that this hypercorrected form was then experiencing an on-going process of diffusion, not exclusively in the written compositions of the learned circles, but more importantly in the everyday speech of those not strictly belonging to that scientific elite, as the more popular text-types actually corroborate. 3.2. Qualitative analysis This section explores the distribution of pleonastic that from a qualitative standpoint to ascertain whether there is a likely preference for a particular type of subordinator, both across time and text types. Figure 2 therefore reproduces the occurrence of the phenomenon in combination with different subordinators in the periods under scrutiny, where the figures have also been normalized for the sake for comparison. The subordinators after that, as that, because that, but that, for that and if that have been chosen in view of their higher occurrence if compared with the other pleonastic constructions. The results, on the one hand, allows us to formulate that the phenomenon diffuses differently with particular types of subordinating conjunctions. In the late Middle English component, the figures tentatively corroborate that the diffusion of pleonastic that is mostly triggered in combination with conditional clauses introduced by the conjunction if, amounting to 37.48 occurrences in the corpus. Next, this pleonasm is also observed to appear in subordinate clauses of time introduced by after (with 23.79 occurrences), when (with 15.49 instances) and until (with 10.77 instances). Finally, it also finds some room 235,43 260,53 15,95 59,38 65,69 55,74 46,31 0 50 100 150 200 250 300 THEORETICAL SURGICAL REMEDIES PT EMEMT MEMT 9 introducing subordinate clauses of reason, in particular with the subordinator for amounting to 11.9 occurrences in the corpus, the subordinator because still negligible in this period. The early modern component, in turn, shows a different state of affairs. The diffusion of pleonastic that in this period substantially decreases in combination with the conditional subordinator if, and the phenomenon is observed to be triggered in subordinate clauses of reason, for again preferred over because in this environments, together with the conjunctions as and after. In addition to this, there is also a widespread use of the phenomenon with other conjunctions, this being the case of the coordinate conjunctions but and or and after whwords, where and when in particular. The relativizers which and who, in turn, are observed to be sporadically used with this pleonasm in this period. Our results also tentatively allow us to date the decline of pleonastic that in the light of the development of particular subordinators across time. Figure 2 reproduces the distribution of the constructions after that, as that, because that, but that, for that and if that over time according to the five sub-periods proposed by MEMT and EMEMT, where the figures have been normalized to a text of 10,000 words for the sake of comparison These subordinators have been chosen in view of their higher occurrence with this pleonasm in the history of English, especially if compared with the others observed to have a more erratic component. After the spread of the phenomenon in the Middle English period, setting aside some particular cases, the results tentatively confirm that the use of pleonastic that in the early modern English period is actively diffusing until the end of the 16th century, as confirmed with the distribution of after that, because that, but that, for that and if that. The beginning of the 17th century witnesses the progressive obliteration of the construction to such an extent that it is considerably reduced to less than five occurrences in cases such as after that (3.24 instances), because that (3.98 instances), but that (4.98 instances) and if that (1.99 instances).