1 The Norman French Involve –ment in Middle English Loss and acquisition of derivational affixes throughout England BA Thesis English Language and Culture, Utrecht University Lennart van der Velden, 3585344 27 June 2013 First reader: Dr. Anita Auer Second reader: Dr. Marcelle Cole 11,022 words 2 1. Introduction The Norman Conquest, led by the Norman French duke William II, introduced the Norman French language to England, although contact between France and England was already prevalent before this. The language was primarily used by royalty and nobility in England after the Norman Conquest. The introduction of Norman French in England and its influence on the English language has raised the question of whether a process of creolisation or language shift or has taken place in this period, which lasted from 1066 to 1500 (e.g. Bailey & Maroldt, Thomason & Kaufman, Dalton-Puffer Creole). Creolisation, as defined by Salikoko Mufwene, occurs when two or more pre-existing varieties enter into a feature pool together, comparable to a gene pool in biology. Similarly to a gene pool, this feature pool contains all elements of the languages involved, and in it “…features associated with the same or similar grammatical functions come to compete with each other” (Mufwene 4). If two or more languages enter into a feature pool together, the result of such a selection procedure is called a creole in cases of intense contact, when the variety has to take on numerous communicative tasks. In cases of more casual contact, where the variety only has to fulfil one purpose (e.g. trade), the variety will be limited. This is called a pidgin. Finally, intense contact between dialects can have competition of features as a result too. The resulting dialect of such competition is called a koiné (4-6). The selection procedure has an immediate connection to the status and number of speakers using the competing varieties. The important difference between this and language shift is that languages engage with one another, whereas in a context of language shift one language has attained a superior position over another to such an extent that the subordinate language declines in numbers of speakers and loses features without replacement. On the topic of the debate regarding Middle English, Christiane Dalton-Puffer claimed that the Middle English language under the yolk of the Norman French displayed symptoms 3 of a recessive language rather than a creolising one. Among other arguments she brings to the table to bolster her claim, she notes that the number of native derivational affixes dropped from 12 to 6 in the Middle English period (Change 46), which indicates a loss of wordformation rules in the language. Derivational affixes are affixes which can change the meaning and class of a word they are attached to. For instance, when the suffix –ment is attached to a verb, e.g. “involve,” creating the noun “involvement,” both the class and the meaning of the word are changed. The supposition that the decrease in derivational affixes in Middle English is symptomatic of language shift, the receding of a language, is backed by Dressler, who regards loss of such word-formation rules as constituting a symptom of language shift, the receding of a language, as it “results from the fact that the language of technology, culture, fashion, etc. has changed from the recessive to the dominant language, at least for the vanguard of speakers most likely to be responsible for creating, adapting and sanctioning neologisms” (qtd. in Dalton-Puffer 45). Note that this statement was made about languages in general, not Middle English specifically. Some of the terms Dressler used as examples not seem appropriate for the Middle English period, but the contention of his words is clear: novelties will be coined in the dominant language. On the topic of affixes being introduced into Middle English from Norman French, Dalton-Puffer notes: “[t]o my mind, they have a lot to do with the recovery of English during the period 1350-1420…” (47). She goes on to present a table showing the amount of tokens and types she has found for a number of Romance suffixes from the entire Middle English period. This data was presented diachronically (showing change over time) in her doctoral dissertation (Influence). To examine whether creolisation or language shift occurred within Middle English on the basis of derivational suffixes, it seems pertinent to carry out an in-depth study into how the loss of word-formation rules and the introduction of the new Norman French affixes progressed for each dialect separately. As this has not yet been done, the present study will take the first step 4 in this direction. In her dissertation Dalton-Puffer does state that “what would now be of interest are case studies digging deeper into the dialectal and stylistic distribution of the phenomena under discussion” (Influence 22). The results of this study are compared to those in Dalton-Puffer’s doctorate, to see whether or not the difference between her results and those for each dialect individually is significant. The importance of such dialectal differentiation is that these different dialects represent different sociolinguistic situations, in which the trends of language change may be different. If the process of loss of native wordformation rules can be shown to progress more quickly than the integration of new rules, it may be argued that that may signal language shift rather than creolisation. If affixes lose their productivity but are concomitantly replaced by Norman French affixes, this may be called creolisation. Based on these two definitions, it may well be possible that even though a process of creolisation occurred in one dialect of the language, a process of shift may have been happening in another, since all these different dialects may be said represent different feature pools. Examining whether or not this is true, whether they are independent enough to warrant this type of research, is another aim of this thesis. Before any research on this can begin, however, the significance of derivational affixes must be explained in lieu of the fact that they are considered mere lexical items (e.g. by Hopper & Traugott). This is discussed in section two. After that, the historical presence of the Norman French language in and around the six dialect regions during the Middle English period is analysed, on the basis of which predictions on the outcome of the corpus study are made. The results of the corpus study carried out to establish rates of integration of Romance affixes and loss of productivity of derivational affixes native to Middle English will be given in section four. Finally, the results will be discussed to show whether we can speak of creolisation or language shift in each variety, and to account for the differences on the basis of population and/or phonetic distance. 5 2.0 The significance of research on derivational affixes An objection that could be raised against the use of derivational affixes in historical linguistic research is that that these items can be considered lexical in nature. Lexical items are easily transferred from one language to the other. If derivational affixes are in fact purely lexical, then what is the difference between work proposed here on endings like –ment and –ity, and research on the integration pattern of any other word, such as “mutton” or “bureau?” It may be argued that research on purely lexical introduction is most fruitful quantitavely (e.g. Dekeyser (qtd. in Dalton-Puffer, Influence 12)). For the present study to have any relevance, then, it is important to show in which ways derivational affixes are not simply lexical. This viewpoint on word-formation rules stems from the fact that, as Hopper and Traugott explain: “[t]he effect seems to be primarily on the lexicon, not the grammar…” (58). They are stored separately in the lexicon and only affect lexicon, hence they are considered to be lexical items. As languages are almost invariably in contact with one another and lexical borrowing is a common occurrence, short of critically examining Dalton-Puffer’s claims of language shift, this study would thus seem to be an inadequate tool to shed light on anything else, e.g. social situation. However, there are many ways in which the status of derivational affixes as lexical items seem to differ from that of others. This is because the function of an item is not the only factor to take into account. For instance, the productivity of a word formation element is considered by many to be more indicative than the actual role an element plays of its grammaticality. Bauer, for instance, states that “[t]he converse of productivity… is ‘lexicalisation’” (qtd. in Brinton & Traugott 17). This position is based on the observation that “in a language like English, the most productive… items are grammatical… whereas the least productive items are certain lexical formatives” (Brinton & Traugott 16-7). As an example of this, Brinton and Traugott present the debate that surrounds the manner adverbial –ly, which 6 concerns whether or not to consider it a derivational affix or an inflection. An inflection is an affix that is required by the grammatical structure of a language in a certain context. For example: “he writes” is a context requiring an –s to be affixed to the verb for agreement with the third person pronoun before it. Derivational affixes are not required, however. In a sentence like “Mike is a jogger” we might say that the –er is obligatory, as the sentence would otherwise not be semantically coherent, but –er in itself is not required to make a word into a noun (Brinton & Traugott 35). An inflection also does not compete with other features similar in meaning, give or take a few unproductive exceptions (such as the plural –s, to which “oxen” as the plural of “ox” is an exception), whereas there is competition in derivational affixation. The affix –ness, for instance, makes a noun out of an adjective, but so does -ity (Brinton & Traugott 36). In studying how to categorise the adverbial –ly, Nevalainen mentions that Marchand excluded –ly from his paper on adverbial affixes, on the basis that it was “fully productive, and thus part of the inflectional system of the language” (qtd. in Nevalainen 147). Although her own conclusion is different, namely that it is “becoming grammaticalised” (Brinton & Traugott 132), this highlights the importance of productivity and (the lack of) competition to the grammatical status of an affix. What is important to take away from this section is that the position of affixes as either grammatical or lexical items is to a large extent contingent upon the way in which they are used. Fossilised affixes, i.e. affixes that only occur in the language as part of certain words, but which are not productive anymore (e.g. “god-head” is fossilised), can on the basis of this definition of grammaticality be considered completely lexical, whereas productive affixes (e.g. establish-ment) are, in fact, grammatical and thus a fruitful basis for research. 3.0 Norman French in and around England This section will present what is currently known and debated about the presence of Norman French speakers in England in the Middle English period. From this information, predictions 7 on the outcome of the dialecto-diachronic corpus study are made. The validity or falsehood of these claims can in turn be tested by the results of the corpus study. The first part of this section presents information on the presence of the Norman French language and people within England, which is presented within the timeframe employed by the Helsinki Corpus, a corpus containing many texts in a variety of styles and genres from every period in the English history1. After that, the presence in the countries neighbouring England will be discussed. This information may help shed light on the linguistic ecology on the Isles, which ought to help us determine beforehand which dialects would be more intensely influenced by the Norman French language. 3.1 Norman French in England 3.1.1 During the Period 1150 – 1250 It was in London that William the Conqueror was crowned king and where he established his court after the Norman Conquest in 1066. The area that was linguistically influenced directly by this sudden influx of speakers of Norman French according to William Rothwell (qtd. in Lusignan 22), spanned ca. 100 miles around London. This 100-mile radius will be taken to be an area in which the Norman French represented a comparatively large proportion of the population. However, when William the Conqueror rose to power, he did not leave the nobility whose monarch he had dethroned to rule with him anywhere in the country. In fact, throughout the country many elite groups were replaced: “[N]ew abbots and bishops were installed… New landlords and a new landowning class appeared” (Gray 3). The old nobility was wiped out by William in an effort to repress rebellion against his reign in a series of military campaigns. These campaigns led to the fact that “[i]n 1072 only one of the twelve earls in England was an Englishman, and he was executed four years later” (Baugh & Cable 1 For more information about the Helsinki Corpus of English texts, visit the manual to the corpus using the following link: http://icame.uib.no/hc/ 8 101). The higher nobility put in place anglicised quickly, however, as a consequence of which “the original Norman monoglots had mostly shifted to English by the late twelfth century” (Machan 74) effectively making Norman French their second language, and that of “other frequenters of the court” (Lusignan 20), such as rural gentry. This anglicisation process may also have been spurred on by the fact that the Norman army in the Conquest did not only consist of Normans, but also of Fleming and Breton soldiers (Gray 4). It may be argued that among them English was considered a suitable lingua franca, a language used, usually as a second language, to communicate with speakers of other languages. The Norman French language remained the language of the kings and queens, however, and thus situated itself as the language of the domain of royalty and nobility. However, Norman French was spread as a second language by the ones indirectly involved in the court, such as the rural gentry and the urban merchant elite as a in order to make trade with the new ruling class possible (Lusignan 20). This situation, as Tim William Machan sees it, meant that “contact between English and French… Should be erased from England’s linguistic ecology” (81). Instead, he proposes that the influence of bilingualism is what accounts for the influence of Norman French on English. This bilingualism not only occurred in English upper-class people learning Norman French as a necessity for contact with the royal court and nobility, but also in Norman French lowerclass henchmen of the new Norman nobility: “[I]t is reasonable to assume that a French soldier settled on a manor with a few hundred English peasants would soon learn the language of the people among whom his lot was cast” (Baugh & Cable 104). Baugh & Cable do not conjecture knowledge of both languages to be limited to these two groups either, however, as they also conjecture that “a knowledge of English was not uncommon at the end of the twelfth century among those who habitually used French; that among churchmen and men of education it was even to be expected; and that among those whose activities brought them into 9 contact with both upper and lower classes the ability to speak both languages was quite general” (112). In the field of religion, Latin had always been predominant, and in the fields where this was traditionally used, it was not challenged by Norman French or English (Momma & Matto 84), although its status from a Norman perspective had changed due to contact with English culture. In England, the English language was frequently used for writing, which also happened with French in England. According to Ian Short this was “‘one of the most important, and one of the least widely recognized, aspects of the new intellectual vitality of the twelfth century” (qtd. in Tyler 165). At any rate, at the end of this period a number of crucial socio-political events took place that put a strain on the connection between England and mainland France. For one, King John of England lost the territory of Normandy, of which he had been duke before, to King Philip of France (Baugh & Cable 116-7). A consequence of this was that “[k]ing and nobles were now forced to look upon England as their first concern” (117). As this concerned the king’s property on the continent, the members of the nobility who also owned land on both the continent and England, and for whom “it might be difficult … to say whether they belonged more to England or to the continent” (117) had to decide to whom they were loyal. This became increasingly important to them when “…the king of France announced that he had confiscated the lands of several great barons, including the earls of Warenne, Arundel, Leicester, and Clare, and of all those knights who had their abode in England” (117) in the same year. This forced many to choose which property they preferred to retain. King John forced similar decisions to be made, resulting in most of the nobility severing ties with either side, and although some managed to keep land on both sides, “double allegiance was generally felt to be awkward” (118). In 1244 the matter was wholly resolved by the king of France, who proclaimed as follows: “‘As it is impossible that any man living in my kingdom, 10 and having possessions in England, can competently serve two masters, he must either inseparably attach himself to me or to the king of England’” (Paris, qtd. in Baugh & Cable 118). As a consequence of this, by 1250 the nobility of England can be said only to have ruled English land (119). Baugh & Cable, on the basis of these facts, draw the conclusion that “in the thirteenth and fourteenth centuries… maintenance [of the use of French] became increasingly artificial” (116). The revival of English might have been halted by the large numbers of immigrants from France (121-2) and the social status of French as a literary language throughout continental Europe (122-3). The English language, they state, would start to retrieve what ground it lost around this time. As will be shown, this contradicts evidence that the peak of the use of the French langauge was yet to come. 3.1.2 During the Period 1250 – 1350 About the beginning of the period 1250-1350 Baugh & Cable state that “when the separation of the English nobles from their interests in France had been about completed, English was becoming a matter of general use among the upper classes” (123). However, an important development in this period is the professionalisation of the judicial system under Edward I (Lusignan 21), which established the French language as not only a prestigious one, but also as a professional necessity and a requirement for anyone who aspired to a position in the judiciary system (20). Thus the status of the French language seems only to be affirmed instead of declining in use. The judiciary was yet another inroad for the French language to affect English. As James Milroy (qtd. in Lusignan 25) explained, the language used by social groups which provide a cohesive environment and do not interact with other groups tends to resist evolution, but the language of people who find themselves in contact with more than one of such groups does change. To stick to Mufwene’s notion of a feature pool: the features of the languages are mixed in those speakers who use both. In the case of the judicial system, for instance, these intermediaries would be stewards to bailiffs rather than judges who may be 11 assumed to speak French (25). This bilingualism actually maintained the diglossia (social divide between two languages, one being prestige and the other regarded as inferior) that existed within society, because “those commanding several languages were characteristically educated, upper-rank individuals” (Machan 77). The low status of English is highlighted by facts such as that “[t]he English language was considered lowly and inadequate, in fact ‘as grammatically unable to represent truth in the way Latin does’” (76). Apart from that, sources pertaining to the Church, such as the administrative documents of Westminster Abbey from 1283 to 1307, which were written in French (Oliva 96), the Hereford canonisation trial of Thomas Cantilupe in 1307 (Lusignan 20), during which half of the clerics opted to give their deposition in French, and the kitcheners’ accounts of the Campsey Ash Priory, a nunnery of “middling social ranks rather than from the elite groups” (Oliva 98) dating from 1298 to 1303 seem to indicate that the introduction of Norman church leaders may have lead to their language being introduced as the spoken language in some circles. In fact, it has even been asserted (e.g. by Kibbee, qtd. in Oliva 97) that certain female nunneries only allowed entrance to French-speaking girls and women exclusively, because they upheld high entry standards and the French-speaking community was generally the elite of society. It is important to realise, however, that many monks (e.g. Chaucer) wrote predominantly in English, and that French can therefore not be considered the absolute language connected to the domain of monastic writing. Documents dating from this period also indicate the use of French in British towns in certain contexts. According to Britnell, ordinances that were meant for publication also occurred in the French language (84). Similarly, taking oaths to the throne was performed in Norman French. Apart from that, since the entire aristocracy had Norman French as their first language at the time, it was used “as a language of recognised high status for communicating with external authorities such as the king and council” (Britnell 87). This does not mean, 12 however, that everyone who corresponded with external authorities (e.g. guild members) necessarily had any command of the language. Britnell assumes that “it may even be that boroughs had to employ specialist clerks to compose texts in French for them” (83). Strikingly, Baugh & Cable draw the same conclusion as may be drawn from the other data presented, namely one of increased introduction of French words into English. This is because in both scenarios the number of bilinguals increased. However, as they feel that “[i]n the fourteenth century English won its way back into universal use, and in the fifteenth century French all but disappeared” (116), this period must be the one they expect to be the one in which such adoption peaked. In this their accounts differ from other sources, as can be read in section 3.1.3. 3.1.3 During the Period 1350 – 1420 According to some of the aforementioned scholars (Britnell, Lusignan), the use of French in England has peaked between 1350 and 1420. The exact period seems to be debated, however. Britnell (89) asserts that it was the period 1350-1415, and Lusignan similarly states that it must have lasted a “long fourteenth century” (22). If “the Golden Age of the King’s French” (22) lies in this period, then it may be argued that the largest degree of change in use of either Norman-French or English affixes occurred in this period. If, however, Baugh & Cable are correct in asserting 1300 as the year in which English reclaimed its status within society, then less change driven by Norman French should occur in this period than it did in the previous. They do note that “[e]ven at the close of the century it was used in Parliament, in the law courts, in public negotiations generally…” (123) French was used. Later on it is noted, however, that “…English had largely taken its place” in the fourteenth century (134). Whatever the case may be, it is true that “[t]he first time that contemporary records admit that Parliament was conducted in English… is 1362” (Texas, italics removed). They also note the decline in proficiency of French as one of the signals of language shift in favour of English 13 (124), as well as the fact that knowledge of English by nobility was by this time assumed (125). On top of that, on the basis of the popularity of educational treatises teaching children French through the English medium, they state: “[W]e may feel quite sure that the mother tongue of the children of the nobility in the year 1300 was, in many cases, English” (125). Here the validity of Baugh & Cable’s claims may truly be tested. If the French language was a foreign language to most of the nobility at this point, mostly overtaken by English in court and other domains, then the use of derivational affixes originally from French must be comparable to that of the previous period. These affixes would have eeked out a place in the English language in the previous period, which they would have kept in this period. If it is the period in which the use of French peaked, however, this period would arguably show a significant increase in the employment of these affixes, because their place within the feature pool would be firmest. In testing the claims made on the status of French in this period, however, it is important to note that this is also the period in which the Great Vowel Shift started. This development, which became especially important at the beginning of the 15th century, had a great impact on English pronunciation, raising many of its vowels. This development was allegedly spurred by “a popular desire to develop pronunciation – over a generation – towards the socially prestigious East-Midlands dialect & its sounds” (Plymouth). This is also why many more texts have been written in this dialect in this period. It may be that, due to a desire to employ that dialect, the use of affixes may also be mirrored, which could lower frequencies of French affixes occurring in areas in which they had been used considerably more often, or increase their use in areas in which they were employed less often. Whatever the case, the area in which people used East-Midlands English grew considerably in this period. 14 3.1.4 During the Period 1420 – 1500 Lusignan and Britnell, who placed the Golden Age of French somewhere in the previous period, claim it to have ended by 1420. Britnell even states that after the peak the language was “…replaced by English quite rapidly after [1415]” (89). The Great Vowel Shift took most of its effect in the fifteenth and sixteenth century (Harvard)2, i.e. most texts from this period would have been written in the East-Midlands dialect, which has an effect on the composition of the Helinki Corpus of English texts, as discussed below. Concomitantly, William Caxton brought his printing press from the European continent to the British Isles in 1476. Since he was the first person to print books in England, he adapted texts to his spelling. Whether he also altered the affixation of texts is questionable. Since his influence may only be felt in the final few years of this period, however, the results of this study will not be significant to show this effect. This might be studied in future research. 3.2 Norman-French Around England There are a number of indications that the Norman French spread their influence to the other realms on the British Isles. Importantly, “Scotland was an independent realm that frequently allied itself with France against England... In a recent study, Cynthia Neville demonstrates that there was a significant Anglo-Norman emigration into the southern part of Scotland, which has been encouraged by the king from around 1125” (Lusignan 23). This means that, although the Northern has not been shown to have been affected much by the Norman French in England itself, their feature pool might have been invaded by French from the north. Not only that, but Trotter (qtd. in Lusignan 22) also identified a large number of Anglo-Norman texts as coming from Wales. Even here multilingualism seemed to be largely commonplace for the higher rungs of society, as “[i]n many cases, these documents were very probably written by Welsh people themselves, and some of them contained words the origins of which This source specifically states the Great Vowel Shift to “[b]egin… in the twelfth century and continuing until the eighteenth century (but with its main effects in the fifteenth and early sixteenth centuries)” (Harvard). 2 15 were Celtic” (22). In Ireland, too, the French language “was used in Irish administrative and literary documents between the thirteenth and fifteenth centuries” (22). Based on these facts, some predictions regarding the outcome of the study in this thesis can be drawn. For one, since both Ireland and Wales are to the west of the country, it may be speculated that if the presence of Anglo-Normans in those areas was more significant than in the West-Midlands at the dates given, this dialect might therefore be affected by the French language more intensely. Likewise, since a large proportion of the population along the Scottish border was Norman French, this might have affected the Northern Middle English dialect to integrate features of Norman French more rapidly. 4.0 Data and Methods 4.1 Data For her doctorate, Dalton-Puffer relied on the Helsinki Corpus of English Texts (Helsinki) exclusively. As stated before, it covers every period in the history of the English language. Not only that, but it also provides useful annotations, providing information such as to which dialect a text has been ascribed. Because Dalton-Puffer used this corpus, its periodisation has also been used for the present study. It is useful for diachronic studies as it contains a relatively large number of texts for each of the periods it distinguishes. In Dalton-Puffer’s case, however, the corpus did not have to be subdivided dialectally as well as diachronically. She did warn on this topic that “the HC is too small to yield reliable results from a process of breaking down the corpus by an increasing number of parameter categories” (Influence 22). This prediction turned out to be true, as the result of dividing the corpus up according to dialect as well as time period yielded an unequal number of texts for each dialect per period, as can be shown in Figure 1. The number of texts has been calculated using the number of times the designers of the corpus saw fit to reference a part of the text, instead of using the number of individual manuscripts. 16 Dialect E-Midlands Kentish Northern Southern W-Midlands ME I 6 6 5 15 ME II 6 2 7 7 ME III 37 4 6 4 ME IV 30 8 5 3 Fig. 1 - Dialectal differentiation Helsinki Corpus Texts From this table it becomes very clear that some dialects are overrepresented, whereas others almost never appear. The fact that there is such a large number of East-Midlands texts has to do with the fact that “[b]y the second half of the 14th century the dialect of London and the area immediately to the northeast, which had once been Kentish, was thoroughly EastMidland…” (Pennsylvania). Authors like Chaucer and the printer Caxton, then, also fell under this dialect. This may have skewed Dalton-Puffer’s results to resemble only the integration/loss pattern of the East-Midlands dialect.3 Because of this discrepancy between dialects in the Helsinki Corpus, to perform this study, the corpus had to be expanded. This has been done by using resources found in several other corpora, such as the Seville Corpus of Northern English (Seville), which contains northern English texts from the 7th century to the 16th. A full list of sources may be found in the Works Cited section. Including these extra texts yielded the numbers shown in Figure 2 below. Dialect E-Midlands Kentish Northern Southern W-Midlands ME I 7 6 3 5 15 ME II 6 2 7 7 7 ME III 37 4 6 6 4 ME IV 30 8 5 3 Fig. 2 - Corpus expanded for dialectal research 3 The dialectal differentiation has been based on the COCOA information given in the Helsinki Corpus. Also, the texts added to the corpus to broaden it have been categorised on the basis of other researchers’ findings. Factors that could have influenced the apparent dialect of the texts (e.g. scribes copying texts into their own dialect) and their effects on the texts have been left to the discretion of those who have categorised the texts and are not taken into account. 17 It is easy to see that the corpus as it presently stands still has its imbalances, but at least almost every dialect is represented with a number of texts. The number of words per subcorpus will be presented with the results of the study. The discrepancy in the sizes of subcorpora do not directly influence the results either, as they will be calculated as words per 1,000. This will give us a percentual number, not an absolute one. The disadvantage of having fewer texts is then that we cannot be as sure of the results. Apparently, no work written in the Kentish dialect from the fourth period has been preserved. The four texts in the third period have not been taken from regions northeast of London, but southeast of it. The genre of the texts added to the corpus was varied, in line with the structure of the Helsinki Corpus, as “an attempt was made to include a wide variety of different text types” (Dalton-Puffer 19) in its construction. The added texts include testaments, fiction and hymns, among other types. 4.2 Methods In her doctorate, Dalton-Puffer worked within the framework of Natural Morphology, as described by Dressler (Morphonology). This model works on the basis of “semiotic transparency” (Dalton-Puffer, Influence 54). It is with the form of the word-formation rule as well as its semantic consequences, and therefore largely based on language-internal factors. It works on the basis of the notions “naturalness” and “markedness.” If a rule is “semiotically better, more transparent,” it will be more natural. Conversely, the less this is the case, the more marked it is. Within this framework, morphemes are given values both on their morphosemantic transparency, which is to say the level to which the morphological change makes clear in what way the meaning of the word has changed. The markedness value is given by determining the place of an affix on a cline, on which the most natural position is given the value 1, whereas the most marked is given a value of 6. The most natural morphological process is simply affixation (e.g “free-dom, accuse-ment” (55)). When this occurs, the word in which the process occurs is not altered in and of itself. The most marked 18 process on this cline, on the other hand, is subtraction (e.g. “Russian logika ‘logic’ : logik ‘logician’” (55)), as for this process the original word is changed, but nothing is added to signal the change in meaning, which is the least transparent way to express a change to the word. Apart from this, there is also a morphotactic (i.e. to do with the phonological effects of putting the word and the affix together) transparency cline which awards a value to morphological processes. This cline focuses on the form of the processed word. On this cline, the most natural process is “intrinsic allophonic phonological rules” (Dalton-Puffer 57), meaning the word is affixed, pronounced the same way, and retains its previous syllable boundaries. However, the most extreme and least transparent morphological process on the cline is that of strong suppletion, in which much, if not all of the word is transformed (e.g. “gall : bilious; am is are be” (Dalton-Puffer 57)). The cognitive approach that Dalton-Puffer employed yields important factors to take into account even when working from a more sociolinguistic creolist perspective as this thesis sets out to do. This approach has been rejected by Dalton-Puffer on the basis of the following, among other things (Influence 48): A) If Middle English is a Creole and B) if a Creole is an extended pidgin C) then there must have existed a pidgin before Middle English. It is recognised that Old English was far from a pidgin, and it certainly did not become more complex during the Middle English period to become suitable for use in more domains. In fact, Old English had complex case, person and even derivational systems which were simplified in this period, which seems to go against this idea of creolisation having taken place. However, when viewed from Mufwene’s perspective, a creole comes into existence through extensive contact between two languages, whereas pidgins “have traditionally been characterised as reduced linguistic systems which are used for specific communicative 19 functions” (7) and are therefore relatively simple. It is true that some pidgins have later taken on more functions, which he calls expanded pidgins rather than creoles, as these originate in settlement colonies rather than trade colonies, which are “marked by contacts that were initially regular and intimate…” (9). The difference between a pidgin and a creole, then, lies in their origination, and one does not follow from the other. Consequently, from this point of view the second premise of Dalton-Puffer’s deduction is not accurate, so approaching Middle English from a creolist perspective is not problematic for this reason. She would probably argue that this makes the difference between creolisation and language contact (49) in light of this view. However, Mufwene argued in the case of koinésation, creolisation and pidginisation that “the outcomes show apparent replacive adoptions by some dialects of elements from other dialects, more like the result of competition and selection than any kind of common denominators of the dialects in contact” (5). While Mufwene stated this with regard to dialects, but it can be used to draw a useful distinction between creolisation and other forms of language contact, namely in the replaciveness of the contact between languages. This would not happen in cases of (non-basic) word borrowing. It is also important to consider creolisation as a process which can stop at any point and on any scale rather than one that must inevitably lead to a creole. Nevertheless, in taking the creolist approach it is important to bear in mind the results of Dalton-Puffer’s research, so that from a Natural Morphology perspective, the compared affixes are not too dissimilar. She took into account both the naturalness of the affixes (focusing in this regard on morphotactic transparency), and the semantic functions that they perform. She concludes that “[i]n the end it became fairly clear that semantics was just as powerful a driving force behind the shifts and reshuffles in the Middle English derivational system [as naturalness]” (227). Seeing as the aim of this paper is to look at socia-political and geographical factors influencing these shifts, both these factors should be neutralised, i.e. 20 should have the least amount of influence possible. This can only be done by comparing affixes which are as close to one another in these respects as possible. For these reasons, -DOM and –ITE have been chosen as nominal affixes, and –FUL and –ABLE have been chosen as adjectival affixes, because the Germanic (-DOM, -FUL) and the Romance affix (ITE, -ABLE) in both these sets are similar both in semantic use and in transparency, as is discussed below. 4.2.1 The Affixes –DOM and –ITE These affixes, throughout the Middle English period appeared according to the pattern shown in figures 3 and 4 below. The upper number in each box represents the absolute number of occurrences, and the bracketed number below shows the number of occurrences per 1,000. The number of tokens describes the absolute number of words in which the affix occurs, whereas the number of types is the number of different words made with it. Although the number of tokens (the absolute number of occurrences) in the third period is higher than in the second period, the number of types (i.e. the different forms the affix was attached to) steadily decreased. Conversely, the suffix –ite steadily gained productivity within the language (shown in figure 3). Note that the total number of types does not match the sum of the numbers of each period. This is because the same type can be counted as a type in two or more periods, but will be counted as only one in the total figure. DOM tokens types Size C. (token) Size C. (type) ME I 129 (1.41) 18 (1.09) 113,010 16,555 ME II 41 (0.42) 10 (0.86) 97,480 11,602 ME III ALL 70 (0.38) 9 (0.57) 184,230 15,880 Fig. 3 - Frequency of DOM (Dalton-Puffer 76, corpus size data and words per 1,000 added) 240 23 21 ITE tokens ME I types Size C. (token) Size C. (type) 12 (0.11) 7 (0.42) 113,010 16,555 ME II 57 (0.58) 20 (1.72) 97,480 11,602 ME III 365 (1.98) 71 (4.47) 184,230 15,880 ALL 434 74 Fig. 4 - Frequency of ITE (Dalton-Puffer 106, corpus size data and words per 1,000 added) Not only that, but in terms of transparency these two forms are also similar to one another. –DOM has a score of 1 in 83.3% of the instances in which it occurred (Dalton-Puffer, Influence 110), which corresponds most to the score of 1.5 in 51% of the occurrences of –ITE. This may seem strange, but one needs to bear in mind that 40.5% the occurrences of ITE were unanalysable. Discarding these, the percentage of ones in this category would be 85.7% (118). Semantically, too, the affixes are most similar (Dalton-Puffer 122-3). Both can function to signal collectiveness, i.e. being a part of group N, N being the noun to which the suffix is attached (martyrdom, nobility). On top of that, both can be used to make a noun out of an adjective signalling the noun has property A (e.g. hostility, freedom). These two, then, appear to be ideal for comparison. Since they are almost identical in transparency and semantics, the most important factor for their decline or increase in productivity must be the users themselves. An aspect of –DOM which possibly deserved more attention in Dalton-Puffer’s doctorate, however, is that the time in which –DOM grammaticalised is hard to pinpoint. It was used as a compound in the Old English period. Compounds are two words put together and as such compounds with the same word are rarer than words with the same affix. Among Old English these compounds were OE “wisdom” and “kingdom” (Ringe & Eska 167), which were some of the most commonly found words in Dalton-Puffer’s data (76). On this topic, Dalton-Puffer simply states: “As an independent lexical item in Old English, it meant ‘jurisdiction, state, statute’… To a certain extent the formations [made using –DOM] have 22 preserved this flavour, which is why there are very few denominal doublets with HEDE” (77). Although her assessment of the Old English variant of DOM is correct, this short discussion leaves out the fact that it is unclear when it did grammaticalise, and when, if at all, this occurred in the Middle English period. To examine whether or not Dalton-Puffer was correct in considering –DOM a Middle English suffix, then, would be an extra point of interest for this study. 4.2.2 The Affixes –FUL4 and –ABLE These adjectival adjectives have very similar semantic functions, namely active potential, meaning the thing it is describes “’can V, Vs’” (Dalton-Puffer, Influence 192), the V being the verb to which it is attached. Apart from that, it can describe passive potential, meaning “‘can be V-ed’” (192). Apart from that, both also have a possessional meaning, giving the meaning “having N” to the thing it describes. However, the Germanic -FUL also has a similitudinal (“resembling N”) function which the Romance –ABLE does not share. Still, these two adjectivals are the most similar adjectival affixes described. ABLE tokens types Size C. (token) Size C. (type) ME1 ME2 ME3 ALL 1 (0.01) 1 (0.06) 10 (0.10) 7 (0.60) 176 (0.96) 46 (2.90) 113,010 16,555 97,480 11,602 184,230 15,880 187 48 Fig. 5 - Frequency of ABLE (Dalton-Puffer 183, corpus size data and words per 1,000 added) 4 -FUL as an affix is also not simply an Old English affix transferred to Middle English. In OE it was simply used adjectivally as “full,” although semantically it was richer in that period (Etymology). It was used for compounding (OE “cupfull” “handfull”). Dalton-Puffer states, however, that “the de-adjectival and deverbal FUL-formations in the corpus seem to mark the onset of a widening of the morphological possibilities of the suffix…” (169). As de-adjectival suffixes already appear in the first Middle English period, the word may be said to have grammaticalised to a fully functional Middle English affix from the onset. 23 FUL tokens ME1 types Size C. (token) Size C. (type) 155 (1.37) 46 (2.78) 113,010 16,555 ME2 ME3 95 (0.97) 22 (1.90) 97,480 11,602 ALL 215 (1.17) 37 (2.33) 184,230 15,880 465 78 Fig. 6 - Frequency of FUL (Dalton-Puffer 169, corpus size data and words per 1,000 added) As expected, -ABLE follows a steady pattern of increased productivity, as is shown in figure 4. –FUL, however, does not consistently decline. From the first period to the second, the affix seems to lose a lot of productivity, coming back with a vengeance in the third period. This pattern would be interesting to analyse dialectally, as it might show that some dialects do embrace the affix after the second period, whereas some others may not. For the reasons laid out in this section, these two adjectival affixes seem ideal for analysis in this study. The frequency of these suffixes will be analysed using the program Simple Concordance Program 4.0.9 (Concordance), the results of which are compared to the general tendencies found by Dalton-Puffer in the tables above. This will reveal whether or not dialectal differentiation makes a significant difference. As Dalton-Puffer notes: “words with the prefix un- were not counted as separate types if the corpus also contains the positive form” (Influence 168). Similarly, all different spelling types occurring for the same word and affix are considered one type. 4.3 Results Below the results of the investigation into the increase or decrease in productivity of the aforementioned affixes are displayed, analysed and compared to those appearing in DaltonPuffer’s study. The results per affix are first shown in absolute numbers, then their relative rise or decline in number will be presented graphically and compared to that in DaltonPuffer’s doctorate. 24 4.3.1 East Midlands ME Results As can be seen in figure 7, in absolute numbers especially, this table is similar in some regards to the results from Dalton-Puffer’s study. The data for –DOM shows a pattern similar to that in Dalton-Puffer’s data. Although token numbers might increase (as is shown best between period three and four), the number of types stays relatively stable, indicating fossilisation according to Dalton-Puffer: “[T]his suggests that there exists an array of fairly fixed expressions in frequnt use without much spontaneous productivity elsewhere” (76). This need not be due to affixation, however. Rather, some particular compounds may be lost containing the word DOM. Also similar is the way in which the affix –ITE entered into the language. The most dramatic increase in token and type numbers occurred between ME III and ME IV in Dalton-Puffer’s results, which she considered “typical of the Romance suffixes” (Influences 106). However, in this study this leap is found between ME II and ME III. It appears that a decrease in use of –DOM cannot be correlated, as a decline in these numbers can already be seen between ME I and ME II, which does correlate with DaltonPuffer’s suggestion that it had done so before ME II (76). There is a great difference, also, in the numbers regarding the affix FUL, especially in the earlier periods. Dalton-Puffer’s results show a decrease in both token and type numbers between ME I and ME II, whereas the opposite seems to be the case in this dialect. This most definitely defies the analysis of “a fairly stable pattern” (168), however appropriate it seemed for the language as a whole. Strangely, the occurrences of –FUL seem to be in decline slightly in the third period, both in type and token counts. The fourth period did not occur in Dalton-Puffer’s research, but this development does show that the affix may not have been invulnerable to the influx of French affixes. In fact, the moment words with –ABLE make up a significant proportion of words in total, the position of -FUL starts to weaken when considering the normalised figures. 25 DOM Tokens ME I ME II ABLE Tokens 28 (0.66) 10 (1.50) ME I 4 (0.10) 4 (0.60) ME I 27 (0.64) 8 (1.20) ME I - Types - Types ITE Tokens Types FUL Tokens Types Size C. (token) Size C. (type) 42,154 6,656 18 (0.36) 6 (0.84) ME II 17 (0.34) 17 (2.37) ME II 115 (2.27) 23 (3.21) ME II 6 (0.12) 3 (0.42) 50,693 7,163 ME III 30 (0.25) 8 (0.61) ME III 231 (1.94) 54 (4.15) ME III 210 (1.76) 23 (1.77) ME III 153 (1.28) 49 (3.76) 119.370 13,021 ME IV 124 (0.96) 9 (0.53) ME IV 264 (2.05) 51 (2.99) ME IV 142 (1.10) 41 (1.23) ME IV 145 (1.13) 52 (3.04) 128.630 17,081 Fig. 7 - East Midlands Distribution Affixes 4.3.2 Kentish Results Before examining the results of the corpus study on Kentish, it must be said that the subcorpus used for this dialect is significantly smaller than that of any other dialect. Moreover, no text at all has been found for the fourth Middle English period. It must be stressed therefore that any tendencies found in the analysis of the results are nothing more than an indication. Texts in this dialect are rare, among other things because “no major literary texts” (Standard) have been written in this dialect. If, at a later date, a more complete corpus is compiled, this study could be performed with a higher degree of certainty. With this in mind, it seems unfair to take the absence of a particular affix at face value. Rather, it would be more appropriate to see whether or not this absence can be said to be anything less than expected, by calculating how many occurrences per 1,000 words it would take for one token to appear. The closest we can get to an estimation of how often the affix occurred is “fewer than x per 1,000 words.” If this number is smaller than that found in 26 Dalton-Puffer’s number, then that might be considered significant, because then the absence of the affix in a body of text of this size might be called unexpected. The results, then, show numbers that are difficult to analyse. In absolute numbers, -DOM types show stability again, regardless of the number of tokens, which is expected. –ITE in ME I did not occur once, but this is also unsurprising, as it occurred in 0.11 words out of 1,000, and the only estimation we can make on the basis of this corpus is that its presence is smaller than 0.16 out of 1,000 words. –ITE, then, shows a dramatic increase and then a drop. Again, the corpus for both these periods is very small, but as a similar pattern also appears in –ABLE, the other French affix, it may be said that what remained Kentish after the East-Midlands dialect area expanded was more Germanic than the rest. This is also shown by the fact that –FUL as an affix does seem relatively stable in this dialect. Whatever the case is, it cannot be claimed based on this data that Kentish, due to its proximity to the East Midlands has been influenced to incorporate the French affixes significantly quickly. DOM Tokens Types ITE Tokens Types FUL Tokens Types ABLE Tokens Types Size C. (token) Size C. (type) ME I ME II 6 (0.94) 2 (1.13) ME I (>0.16) ME I 9 (1.42) 3 (0.56) ME I (>0.16) 6,355 1,770 Fig. 8 - Kentish Distribution Affixes 21 (1.36) 3 (1.21) ME II 24 (1.55) 7 (2.82) ME II 21 (1.36) 14 (3.62) ME II 3 (0.19) 3 (1.21) 15,480 2,483 ME III 2 (0.28) 2 (1.25) ME III 2 (0.28) 1 (0.62) ME III 14 (1.93) 8 (2.50) ME III (>0.14) 7,254 1,605 ME IV ME IV ME IV ME IV - - 27 4.3.3 Northern ME Results In the subcorpus compiled for this dialect, again, two periods have been represented by very few words, each consisting of few words as well. The first period is especially poorly represented, which is for the most part due to the scarcity of texts from this period hailing from the regions where this dialect was spoken. It is with extreme caution, therefore, that conclusions must be drawn from this data. The number of tokens per 1,000 of –DOM comes very close to what would be expected from Dalton-Puffer’s data in the third period, to decline afterwards. -ITE again shows the greatest increase in productity from the second period to the third, increasing even further later on. On the basis of this corpus, it is hard to say whether or not –DOM fossilised, if at all, and what sort of influence –ITE might have had on it. –FUL, yet again, does show a reasonably stable pattern, increasing in productivity consistently. –ABLE does not seem to be able to get a steady foothold in this dialect. After a modest spike in the third period, the affix seems to have lost ground by the fourth period. This possibly assured the continued stability of –FUL, where its productivity declined in the East-Midlands. DOM Tokens ME I 1 (1.21) 1 (2.60) Types ITE Tokens ME I Types (>1.21) - FUL Tokens ME I - Types - ABLE Tokens ME I - Types - Size C. (token) Size C. (type) 821 385 Fig. 9 - Northern Derivational Affixes ME II ME III ME IV 6 (2.10) 4 (3.87) ME II 6 (2.10) 4 (3.87) ME II 3 (1.05) 2 (1.94) ME II (>0.35) - 10 (0.36) 5 (1.15) ME III 21 (0.77) 13 (3.00) ME III 33 (1.20) 15 (3.23) ME III 4 (0.15) 3 (0.69) 27,445 4,337 8 (0.23) 2 (0.33) ME IV 37 (1.06) 14 (2.28) ME IV 56 (1.60) 26 (2.77) ME IV 1 (0.03) 1 (0.16) 35,032 6,142 2,854 1,033 28 4.3.4 Southern ME Results In the Southern dialect, again, we see a clear pattern of apparent fossilisation occurring in –DOM, the type count not increasing per 1,000 when the number of tokens does. The integration pattern of –ITE also roughly corresponds to that in Dalton-Puffer’s data, the largest increase being between the second and third period. However, at every stage the number of occurrences per 1,000 words stays below Dalton-Puffer’s findings. The numbers of this affix per 1,000 are consistently lower than those in her data, which is clearest in ME III, where it is little more than half. Similarly, -ABLE, only shows very slight integration allround, occurring more than three times as often per 1,000 words in Dalton-Puffer’s study. What is more, like the Northern dialect, this affix seems to have decreased in productivity from the third to the fourth period. –FUL, however, does not show a stable pattern in this dialect. Similar to Dalton-Puffer’s findings, there is a dip in productivity in the second period, from which the affix seems to recover. Suprisingly, affix use allround increases in the fourth period. Overall, it seems that DOM has apparently fossilised separately from the influence of Norman French, as the number of tokens for this affix increased when this too increased for – ITE. Furthermore, the productivity both Norman French affixes in this dialect do not seem to be as high as would have been expected from Dalton-Puffer’s data. Language contact, then, may on the basis of this study be said to be less intense. 29 DOM Tokens ME I ME II 13 (0.60) 6 (1.51) ME I (>0.05) - Types ITE Tokens Types FUL Tokens ABLE Tokens ME I 15 (0.70) 7 (0.75) ME I - Types - Types Size C. (token) Size C. (type) 21,549 3,976 5 (0.15) 4 (0.89) ME II 19 (0.56) 9 (1.99) ME II 9 (0.27) 6 (1.11) ME II 1 (0.03) 1 (0.22) 33,652 4,505 ME III 22 (0.57) 5 (0.86) ME III 46 (1.18) 15 (2.57) ME III 26 (0.67) 19 (2.57) ME III 11 (0.28) 6 (1.03) 38,819 5,834 ME IV 17 (0.87) 3 (0.78) ME IV 50 (2.55) 15 (3.89) ME IV 15 (0.76) 10 (2.08) ME IV 9 (0.46) 5 (1.30) 19,626 3,853 Fig. 10 – Southern Derivational Affixes 4.3.5 West Midlands ME Results In the West Midlands dialect the decreased productivity of –DOM occurring at the onset of the Middle English period is clearest due to the relatively large number of types occurring. This shows a more conservative Germanic pattern than can be found in other dialects. Again, the number of types per 1,000 types decreases throughout the periods, but in this corpus, so do the numbers of tokens. Sadly, the subcorpus used for analysing this dialect is small in the second period, so this can, again, only be used as an indication. Strangely, the number of types and tokens for –ITE are large from the onset, only grow very slightly into the second period, spike in the third, and decrease in the fourth period. This might indicate very early contact with the Norman French, especially since the affix –ABLE only occurs in this dialect in this period. They do not, however, increase the way expected for a dialect in intense contact with French, especially considering the numbers in the fourth Middle English period. For – 30 FUL, the numbers are significantly lower, even per 1,000 words, but the pattern is roughly the same, showing a dip in the second period, then climbing up again. What we can glean from this is that there has been a process of creolisation in the first Middle English period, but that from then on (although we cannot be sure of the second period), the dialect shows signs of shift, most clearly visible in the fourth period, in which very few tokens were gathered. DOM Tokens ME I Types ITE Tokens Types FUL Tokens Types ABLE Tokens Types Size C. (token) Size C. (type) ME II 84 (1.32) 14 (1.50) ME I 16 (0.25) 9 (0.96) ME I 121 (1.90) 49 (1.39) ME I 1 (0.02) 1 (0.11) 63,520 9,327 ME III 3 (0.59) 2 (1.21) ME II 3 (0.59) 3 (1.83) ME II 2 (0.39) 2 (1.21) ME II (>0.20) 5,127 1,643 11 (0.62) 4 (1.10) ME III 21 (1.18) 8 (2.21) ME III 15 (0.84) 11 (1.93) ME III 1 (0.56) 1 (0.28) 17,815 3,625 ME IV 5 (0.21) 3 (0.78) ME IV 5 (0.21) 4 (1.04) ME IV 13 (0.55) 9 (1.56) ME IV 1 (0.04) 1 (0.26) 23,727 3,841 Fig. 11 - West Midlands Derivational Affixes 4.3.6 Graphical Representation of Results 2.5 Words/1,000 2 E-Midlands Kentish 1.5 Northern Southern 1 W-Midlands Dalton-Puffer 0.5 0 ME I ME II ME III Fig. 12 – Occurrences –DOM per 1,000 words per dialect Middle English Period Fig. 11 – Occurrences –DOM per 1,000 words per dialect ME IV 31 3 Words/1,000 2.5 E-Midlands 2 Kentish Northern 1.5 Southern W-Midlands 1 Dalton-Puffer 0.5 0 ME I ME II ME III ME IV Middle English Period Fig. 13 – Occurrences –ITE per 1,000 words per dialect 2.5 E-Midlands Kentish 1.5 Northern Southern 1 W-Midlands Dalton-Puffer 0.5 0 ME I ME II ME III ME IV Middle English Period Fig. 14 – Occurrences –FUL per 1,000 words per dialect 1.4 1.2 E-Midlands 1 Words/1,000 Words/1,000 2 Kentish 0.8 Northern 0.6 Southern W-Midlands 0.4 Dalton-Puffer 0.2 0 ME I ME II ME III Middle English Period Fig. 15 – Occurrences –ABLE per 1,000 words per dialect ME IV 32 Figures 12 to 15 above are graphs made on the basis of the occurrences of the studied affixes in normalised figures. In the cases of some affixes, the dialectal differentiation does not seem to have a great impact on the results, such as in –DOM. All of the dialects seem to follow the same pattern with comparable numbers. The two major detractions from the pattern in DaltonPuffer’s results both found in two small corpora, so they may be inaccurate. Others, however, such as –ABLE and –ITE, both newly integrated affixes in the English language at that point, show almost all dialects not following the pattern in Dalton-Puffer’s data. This again indicates that the way in which –DOM evolved may have been largely due to language-internal factors rather than language contact. Another distinct possibility is that –DOM did not behave like a suffix because it was not one yet, in that it had not grammaticalised yet. In Dalton-Puffer’s results, the only reason that –DOM had more tokens to begin with, therefore showing a pattern of decreased productivity, was that the West-Midlands dialect data contained a larger number of them, probably due the fact that in the Old English period West Saxon was a prestigious dialect, and therefore a standard of sorts. This probably caused the West-Midlands dialect to be more conservative in its lexicon of the Old English heritage, hence showing the larger number of tokens. When the French language added to this lexicon, some of the older compounds may have been lost. Since all dialects were affected lexically, as lexicon is relatively easy to integrate, this might have caused the decrease in words containing DOM. At any rate, as –ABLE and –ITE show unstable patterns, it is clear that integration of new affixes has been significantly different for each dialect, indicating that the feature pools of each are separate enough from one another to be considered individually rather than collectively as Dalton-Puffer has done. In the –ABLE graph, the correlation between Dalton-Puffer’s results and those found in the East-Midlands’ subcorpus is shown most clearly. Especially in the ME III period this is unsuprising, as for that period the largest proportion of the Helsinki Corpus consisted of texts written in this dialect. 33 5.0 Discussion of Results Now that the data of the corpus study have been presented, the predictions made on the basis of the historical literature can be compared to it. Initially, it had been predicted that, as the Norman French made London the city of the royal court, the Southern dialect, then still spoken in that area, would be the dialect most affected by Norman French initially; especially considering that the texts in the first period in this study were written roughly a century after their victory in the Norman Conquest. This, however, was not borne out by the results, which rather showed that the West Midlands dialect showed more use of the Norman French affixes in the first Middle English period. The subcorpus for this dialect also had the largest body of texts in this period, most likely to do with the fact that West-Saxon, hailing from the same region, used to be a prestigious Old English dialect. This rich Old English background may, as said before, also serve to account for the large number of types and tokens made with DOM in the first period, because it would likely preserve more linguistic features from that period. Another prediction made was that the presence in the areas surrounding England might have impacted the integration patterns of dialects, specifically the West Midlands and Northern dialects. In the case of the West Midlands, there is some indication that this has been borne out. The fact that –ABLE occurred at all, combined with the fact that –ITE also occurred almost double the average of the dialects calculated by Dalton-Puffer, may hint into this direction. In the case of the Northern dialect, the proximity of Norman French speakers across the borders at first glance does not seem to have impacted the dialect greatly, as both –ABLE and –ITE have not been integrated as quickly as in other dialects. Then there is the discussion between Britnell, Lusignan and Baugh & Cable. On the basis of Baugh & Cable’s claims, who stated that near the beginning of the third period English was the dominant language, it could be predicted that in this period the integration 34 patterns of the new affixes would be largely stabilised, having peaked the period before. For the East-Midlands dialect this appeared to be true, as there were no dramatic changes in their productivity from the third to the fourth period. For the other dialects, however, this does not seem to be the case. This may indicate that facts on which Baugh & Cable based their claims had more of an effect in the East Midlands. This seems to follow, since the Royal court was established in London, where the East-Midlands dialect was spoken. Since the acts by the French king against holding lands both on the continent and on the Isles affected the nobility and royalty most, and these were more prevalent in the East Midlands, they might have had more effect on the language there. Further away, where nobility and certainly royalty made up a smaller portion of the population, these acts would have impacted the dialects less dramatically. Interestingly, however, the increase in productivity of these affixes in the Southern period was highest in the fourth period. This may be explained the same way in which Baugh & Cable explained the increased language contact betweeen English and French when use of French dwindled: the fact that people who formerly used French would have had to learn English afterwards. This may already have happened for nobility, but perhaps not for other groups which used French, such as the immigrants who came to England, after the rapid decline in use of French pointed out by Lusignan. Finally, the largest question remains of whether or not we can see a significant difference between dialects in creolisation or language shift patterns. To assess this, it seems useful to analyse each dialect separately. In doing so, DOM must be kept out of the equation, since based on this dialectal research DOM does not seem to have been an affix in this period. First of all, the East Midlands dialect shows a very clear pattern of integration of the French affixes analysed for this study. This seems to indicate a process of creolisation, since the Germanic affix is seemingly gradually replaced by the Romance affix. Where Kentish is concerned, the corpus size makes it difficult to make any general statements. The most 35 representative part of the corpus is the second period, by which time –ITE seems to have been integrated. -FUL does not seem to lose productivity though the centuries in this dialect, however, whereas –ABLE does not seem to get a foothold. The integration of –ITE without replacement, then, points towards language contact without creolisation or language shift. In the Northern dialect, likewise, –ABLE does not garner a solid position. Use of –ITE is also not as common as it is in other dialects. DOM loses types and tokens, and even seems to be close to loss in the fourth period, but this may mean nothing other than lexical integration. It seems unlikely that this would have been spurred on by –ITE gaining popularity, however. It may signal a process of language shift, however, which may have been caused by the Norman French presence near the region in which this dialect was spoken. In this way, the presence may well have affected the status of English in that area. For the Southern dialect, then, neither creolisation nor language shift seem appropriate. Like the West-Midlands dialect, its integration of –ITE was more gradual than might have been expected on the basis of DaltonPuffer’s data, but its productivity heavily increased from the third to the fourth period. This did not seem to have an effect on the token numbers of DOM, indicating once more that this does not seem to behave like an affix. Numbers of the affix –FUL are low, but relatively stable. Apart from a small drop in the second period the number of occurrences per 1,000 words does not seem to have fluctuated much. The fact that –ABLE was integrated into the dialect, though gradually compared to others, does not seem to have impacted –FUL either. This indicates that, at least for these affixes, competition has not been significant to impact one another’s productivity. Language contact seems to fit this situation rather than language shift or creolisation. Finally, the West-Midlands dialect shows a pattern which leans more towards language shift. –FUL seems to have lost productivity over time, not only from the first to the second period (for which the data was sparse), but also in later periods. At all times after the first period this affix has either the lowest or the second lowest productivity in this 36 dialect of all dialects. The Romance affixes both had a small spike in use in the third period, to be all but lost in the fourth period. In sum, in many ways the dialects seem, on this data, to have shown different integration and loss patterns where these affixes are concerned. Examining the Middle English language as a whole, then, although the results might be helpful in presenting a general picture, does not suffice to show the intricacies of the types of contact that occurred within the dialects that make up the language. Something that should be noted in reading these results is that literacy in the period examined in this thesis was a privilege reserved only for the upper echelons of society. As the Norman French positioned themselves in these circles as well, it may be argued that the language the people who could write would be most strongly influenced by Norman French. Moreover, the style in which these speakers wrote may not represent the way they actually spoke. These results, then, do not necessarily, or even probably do not, represent the varieties spoken by the majority of speakers, who were unable to write. This, however, is the closest a researcher can come to researching the language spoken at that time. As such, it does serve to show the social situation of the time in question. Also, an aspect which may have contributed to the way the different dialects of Middle English were in contact with French, yet which has not been taken into account in this study is the relative phonetic and phonological distance between the dialects of Middle English and Norman French before the Conquest. The dialects will have shifted towards the French language in this regard as well, but it may be argued if this distance was greater, the integration of elements may have been more gradual. For instance, Northern English has been influenced more by Old Norse than other English dialects. If, because of this, the sounds in the language and the rules governing their use were farther removed from those of French, integration from one into the other might have been more difficult because of that. What the 37 effect of these dialectal differences was on the integration patterns of derivational affixes would be interesting for future research. 6. Conclusion The main problem this thesis set out to shed light on was whether or not Middle English showed a pattern of language shift or of creolisation. To do so, the integration and loss patterns of several Romance and Germanic origin affixes have been examined. The study of derivational affixes as markers of language contact had to be defended, however, since derivational affixes are viewed by some as being lexical items. This has been done on the basis of their potential grammaticality. An overview of the current knowledge on the presence of the Norman French and their language throughout the periods that have been examined in this paper has been presented, on the basis of which predictions have been made. These predictions have then been compared to the results of the corpus research done on, and it appeared that some predictions turned out to be more accurate for some dialects than others. This study also provided evidence against DOM being a suffix in the Middle English period, on the basis that its pattern of loss did not correspond to what would be expected from an affix. Apart from that, the results of this corpus research indicate that where in some dialects a pattern of language shift was shown, creolisation appeared to be occurring in others. This shows that the Middle English dialects were separate from one another enough to consider them different feature pools, indicating the importance of taking each dialect separately in researching the Middle English language. 38 WORKS CITED Bailey, Charles J. and Karl Maroldt. “The French Lingeage of English.” Pidgins – Creooles – Languages in Contact. Ed. Jürgen Meisel. Tubinger Beitrage zur Linguistik 75. Tübingen: TBL Verlag-Narr, (1977): 21-53. Print. Baugh, Albert C. and Thomas Cable. A History of the English Language. 5th ed. London: Routledge & Kegan Paul. Marquette University, n.d. Web. 26 Jun. 2013. Brinton, Laurel J. and Elizabeth Closs Traugott. Lexicalization and Language Change. Cambridge: Cambridge University Press, 2005. Print. Britnell, Richard. “Uses of French Language in Medieval English Towns.” Language and Culture of Medieval Britain – The French of England c. 1100 – 1500. Ed. Jocelyn Wogan-Browne. York: York Medieval Press, 2009. 81-89. Print. Dalton-Puffer, Christiane. The French Influence on Middle English Morphology. Berlin: Mouton de Gruyter, 1996. Print. Dalton-Puffer, Christiane. “Middle English is a creole and its opposite: On the value of plausible speculation.” Linguistic Change under Contact Conditions. Ed. Jacek Fisiak. Trends in Linguistics. Studies and Monographs 81 (1995): 35 – 50. Print. “Dialects of Middle English.” The University of Pennsylvania, n.d. Web. 26 Jun. 2013. Dressler, Wolfgang U. Morphonology: The Dynamics of Derivation. USA: Karona Publishers Inc, 1985. Print. “Full.” Online Etymology Dictionary. Douglas Harper, 2011. Web. 28 Jun. 2013. Hopper, Paul J. and Elizabeth Closs Traugott. Grammaticalization. 2nd ed. Eds. P. Austin, J. Bresnan et al. Cambridge: Cambridge University Press, 2003. Print. “Late Middle English” University of Texas at Arlington, n.d. Web. 26 Jun. 2013. Lusignan, Serge. “French Language in Contact with English: Social Context and Linguistic Change (mid-13th–14th centuries).” Language and Culture of Medieval Britain – The 39 French of England c. 1100 – 1500. Ed. Jocelyn Wogan-Browne. York: York Medieval Press, 2009. 19-30. Print. Machan, Tim William. English in the Middle Ages. Oxford: Oxford University Press, 2003. Print. Marchand, Hans. Categories and types of present-day English word-formation. 2nd ed. München: Beck, 1969. Print. “Middle English – Dialects – Caxton & Printing – Emergence of a Standard.” Boston College. The Trustees of Boston College, n.d. Web. 26 Jun. 2013. Milroy, James. Linguistic Variation and Change: On the Historical Sociolinguistic of English. Oxford: Blackwell Publishing Ltd. 1992. Print. Momma, Haruko and Michael Matto, eds. A Companion to the History of the English Language. Hoboken: Wiley-Blackwell, 2008. Print. Mufwene, Salikoko S. The Ecology of Language Evolution. Cambridge: Cambridge University Press, 2001. Print. Nevalainen, Terttu. “The processes of adverb derivation in Late Middle and Early Modern English.” Grammaticalization at Work: Studies of Long-Term Developments in English. Eds. Matti Rissanen, Merja Kytö and Kirsi Heikkonen. Topics in English Linguistics 24. Berlin: Mouton de Gruyter (1997): 145-189. Print. Oliva, Marilyn. “The French of England in Female Convents: The French Kitcheners’ Accounts of Campsey Ash Priory.” Language and Culture of Medieval Britain – The French of England c. 1100 – 1500. Ed. Jocelyn Wogan-Browne. York: York Medieval Press, 2009. 90-102. Print. Paris, Matthew. “Chronica Majora.” Trans. J. A. Giles. Matthew Paris’s English History. Vol. 1. London: Henry J. Bohn, 1854. 481–82. Print. 3 Vols. 40 Reed, Alan. “Simple Concordance Program 4.0.9.” Simple Concordance Program. Alan Reed, 10 Nov. 2012. Web. 26 Jun. 2013. Ringe, Don and Joseph F. Eska. Historical Linguistics: Toward a Twenty-First Century Reintegration. Cambridge: Cambridge University Press, 2013. Print. Short, Ian. “Patrons and Polyglots: French Literature in Twelfth-Century England.” AngloNorman Studies XIV: Proceedings of the Battle Conference 1991. Ed. Marjorie Chibnall. Suffolk: Boydell & Brewer Ltd, 1992. 229 – 250. Print. “The Great Vowel Shift” The Geoffrey Chaucer Page. The President and Fellows of Harvard College, July 27, 2000. Web. 26 Jun. 2013. “The Late Middle-English period (1470 - 1550).” Timeline of the English Language. University of Plymouth, 2007. Web. 26 Jun. 2013. Thomason, Sarah G. and Terrence Kaufman. Language Contact, Creolisation and Genetic Linguistics. Chicago: University of Chicago Press, 1988. Print. Trotter, D. A. “L’anglo-français au pays de galles: une enquête préliminaire.” Revue de Linguistique Romane 58 (1994): 461-487. Print. Tyler, Elizabeth M. “From Old English to Old French.” Language and Culture of Medieval Britain – The French of England c. 1100 – 1500. Ed. Jocelyn Wogan-Browne. York: York Medieval Press, 2009. 164-178. Print. Corpus sources Braswell, Mary Flowers, ed. “Sir Perceval of Galles.” The Camelot Project. University of Rochester, n.d. Web. 26 Jun. 2013. “Cambridge University Library, Add. 6845.” Oxford University Computing Services – University of Oxford Text Archive. University of Oxford, n.d. Web. 26 Jun. 2013. “Carlisle Cumbria ROD/Lons/L Medieval Deeds C1.” Seville Corpus of Northern English. University of Seville, n.d. Web. 27 Jun. 2013. 41 “Cotton, MS Faustina A. v.” Seville Corpus of Northern English. University of Seville, n.d. Web. 27 Jun. 2013. “Cotton, MS.Vespasian A.iii.” Seville Corpus of Northern English. University of Seville, n.d. Web. 27 Jun. 2013. Edinburgh, Royal College of Physicians. “Cursor Mundi.” Seville Corpus of Northern English. University of Seville, n.d. Web. 27 Jun. 2013. Finchdale, Godric. “The Hymns of Saint Godric.” Old and Middle English c. 890 – 1450 – An Anthology. Ed. Elaine Treharne. 3rd ed. West Sussex: Wiley Blackwell, 2010. 324 – 326. Print. “Kingston-upon-Hull Corporation ArchivesBench Book 2 f. 243.” Seville Corpus of Northern English. University of Seville, n.d. Web. 27 Jun. 2013. Kytö, Merja. “Helsinki Corpus of English Texts.” Oxford University Computing Services – University of Oxford Text Archive. University of Oxford, n.d. Web. 26 Jun. 2013. “London, Lambeth Palace Library 216.” The Middle English Grammar Project. University of Stavanger, 16 Mar. 2011. Web. 26 Jun. 2013. “London British Library Cotton Charter iv 18.” Seville Corpus of Northern English. University of Seville, n.d. Web. 27 Jun. 2013. “London British Library Cotton Cleopatra B vi fol. 204v.” Seville Corpus of Northern English. University of Seville, n.d. Web. 27 Jun. 2013. “London, British Library Cotton Julius A v, fols. 180r-181v.” Seville Corpus of Northern English. University of Seville, n.d. Web. 27 Jun. 2013. “Oxford Bodleian Library Bodley 26, fols. 107r-108r.” Seville Corpus of Northern English. University of Seville, n.d. Web. 27 Jun. 2013. “Oxford, Bodleian Library, Top. Kent d.3.” The Middle English Grammar Project. University of Stavanger, 16 Mar. 2011. Web. 26 Jun. 2013. 42 “Oxford, University College 142.” The Middle English Grammar Project. University of Stavanger, 16 Mar. 2011. Web. 26 Jun. 2013. “Rawlinson, MS.C 317.” Seville Corpus of Northern English. University of Seville, n.d. Web. 27 Jun. 2013. “The will of John Smyth of Rolleston.” Seville Corpus of Northern English. University of Seville, n.d. Web. 27 Jun. 2013. 43 VERKLARING: INTELLECTUEEL EIGENDOM De Universiteit Utrecht definieert het verschijnsel “plagiaat” als volgt: Van plagiaat is sprake bij het in een scriptie of ander werkstuk gegevens of tekstgedeelten van anderen overnemen zonder bronvermelding. Onder plagiaat valt onder meer: het knippen en plakken van tekst van digitale bronnen zoals encyclopedieën of digitale tijdschriften zonder aanhalingstekens en verwijzing; het knippen en plakken van teksten van het internet zonder aanhalingstekens en verwijzing; het overnemen van gedrukt materiaal zoals boeken, tijdschriften of encyclopedieën zonder aanhalingstekens of verwijzing; het opnemen van een vertaling van bovengenoemde teksten zonder aanhalingstekens en verwijzing; het parafraseren van bovengenoemde teksten zonder verwijzing. Een parafrase mag nooit bestaan uit louter vervangen van enkele woorden door synoniemen; het overnemen van beeld-, geluids- of testmateriaal van anderen zonder verwijzing en zodoende laten doorgaan voor eigen werk; het overnemen van werk van andere studenten en dit laten doorgaan voor eigen werk. Indien dit gebeurt met toestemming van de andere student is de laatste medeplichtig aan plagiaat; ook wanneer in een gezamenlijk werkstuk door een van de auteurs plagiaat wordt gepleegd, zijn de andere auteurs medeplichtig aan plagiaat, indien zij hadden kunnen of moeten weten dat de ander plagiaat pleegde; het indienen van werkstukken die verworven zijn van een commerciële instelling (zoals een internetsite met uittreksels of papers) of die tegen betaling door iemand anders zijn geschreven. Ik heb de bovenstaande definitie van het verschijnsel “plagiaat” zorgvuldig gelezen, en verklaar hierbij dat ik mij in het aangehechte essay / werkstuk niet schuldig heb gemaakt aan plagiaat. Naam: Studentnummer: Plaats: Datum: Handtekening: