-ed alternate verb forms. A corpus and survey based study

Faculteit Letteren & Wijsbegeerte Hanne De Cock The impact of aspect on -t/-ed alternate verb forms. A corpus and survey based study Masterproef vo...
Author: Corey McDowell
16 downloads 0 Views 972KB Size
Faculteit Letteren & Wijsbegeerte

Hanne De Cock

The impact of aspect on -t/-ed alternate verb forms. A corpus and survey based study

Masterproef voorgedragen tot het behalen van de graad van Master in het Vertalen 2015-2016

Promotor Prof. Dr. De Clerck Vakgroep Vertalen Tolken Communicatie

2

ACKNOWLEDGEMENTS

I could not have written this paper on my own and I would like to use this opportunity to thank everyone who has been there for me during this writing and thinking process. First and foremost, I would like to thank Bernard De Clerck, my supervisor, without whom this dissertation would not have been possible. I have learned so much from his feedback and insights and I am very grateful for the time and effort he put into this paper. I would also like to thank my parents, who have supported me and believed in me without fail during this entire process. Dad, thank you for always being impressed with what I do and what I write, even if I firmly believe it is not my best work. In addition, I owe a big thank you to my Mom for never giving me the idea that I could not do whatever I wanted to do or be whomever I wanted to be. Mom, you are my guidepost for everything. Finally, special gratitude goes to my grandmother for the weekly motivational phone calls during the past four years at University. I am proud to be your grandchild.

3

ABSTRACT

This research explores the impact of aspect on -ed and -t preferences with verbs that have variable inflection. Two types of analysis were carried out: first, corpus data was examined to corroborate the purported link between aspectual durativity/punctuality and -ed/-t preferences as observed in Levin 2009 (a.o.). Secondly, an elicitation experiment with first Bachelor and second Bachelor students at the Department of Applied Language Studies at Ghent University probed into the effect of aspect on verb inflection on L2 learners. The corpus results show a link between -t verb forms and punctuality, but the same could be attested for -ed verb forms. Moreover, the individual results for each verb showed that there is a high level of internal variation. The forcedchoice test among first Bachelor students confirmed our hypothesis. However, in the free-choice test students made different choices and appeared to have a preference for either the -ed or the -t verb form which are not driven by aspectual considerations.

4

TABLE OF CONTENTS 1.

INTRODUCTION .................................................................................................................... 6

2.

THEORETICAL FRAMEWORK ............................................................................................ 8 2.1 Verb regularisation processes and inflectional variation ..................................................... 8 2.2 Impact of variables on -ed and -t selection ........................................................................ 10

3.

2.2.1

Aspect ....................................................................................................................... 10

2.2.2

Transitivity ............................................................................................................... 15

2.2.3

Tense and voice ........................................................................................................ 16

2.2.4

The Principle of Rhythmic Alternation .................................................................... 16

2.2.5

Frequency ................................................................................................................. 20

METHODOLOGY ................................................................................................................. 21 3.1 Corpus data ........................................................................................................................ 21 3.2 Survey ................................................................................................................................ 23

4.

3.2.1

Informants ................................................................................................................ 23

3.2.2

Tests ......................................................................................................................... 23

RESEARCH QUESTIONS .................................................................................................... 26 4.1 Corpus data ........................................................................................................................ 26 4.2 Survey data ........................................................................................................................ 26

5.

DATA ANALYSIS ................................................................................................................. 27 5.1 Corpus data ........................................................................................................................ 27 5.1.1

General results .......................................................................................................... 27

5.1.2

Results individual verbs ........................................................................................... 28

5.2 Survey data ........................................................................................................................ 31 5.2.1

Forced-choice test .................................................................................................... 31

5.2.2

Free-choice test ........................................................................................................ 32

5.2.3

Results individual verbs ........................................................................................... 33

5.2.4

Comparison Bachelor 1 and Bachelor 2 ................................................................... 37

6.

CONCLUSION ....................................................................................................................... 40

7.

BIBLIOGRAPHY ................................................................................................................... 42

5

LIST OF TABLES Table 1 : results forced-choice selection test Quirk (1970) Table 2 : results three-point scale judgment test Quirk (1970) Table 3: number of sentences that was examined in the corpus study Table 4: division punctual and durative sentences Table 5: general results corpus study Table 6: corpus results individual verbs Table 7: attested proportions -ed/-t verb forms GloWbE (De Clerck & Vanopstal) Table 8: results forced-choice test Bachelor 1 Table 9: general results free-choice test Bachelor 1 and 2 Table 10: individual results free-choice test Bachelor 1 and 2 combined Table 11: attested proportions -ed/-t verb forms GloWbE (De Clerck & Vanopstal) Table 12: general outcome free-choice test Bachelor 1 and 2 Table 13: results free-choice test Bachelor 1 and 2

LIST OF APPENDICES Appendix 1: version 1 of the free-choice test Appendix 2: version 2 of the free-choice test Appendix 3: forced-choice test Appendix 4: results corpus study Appendix 5: results free-choice and forced-choice test

6

1. INTRODUCTION There is a limited set of verbs in the English language that take both a regular and an irregular past simple and past participle form. Verbs like burn, learn, dream, spell, smell, for instance, can either take the -ed or -t spelling as in burned/burnt, learned/learnt, dreamt/dreamed, spelt/spelled, smelt/smelled.1 This synchronic variation reflects a diachronic process of verb regularisation (see Liebermann et al. 2007) in which irregular forms are being replaced by their regular counterparts. Some verbs have regularised entirely and the irregular form is now a relic of the past (e.g. helped/helped, laughed/laughed). In other cases, both forms continue to co-exist, though attested proportions of regular and irregular uses may vary substantially across the verbs that belong to this category. De Clerck & Vanopstal (2015), for instance, have shown that dwell shows a strong preference for the irregular –t form across different varieties of English while learn prefers the –ed form. In the latter case, a fraction of –t uses still lingers on and it is still unclear why the low frequent forms do not disappear altogether. One of the explanations could be that the process of regularisation has not been completed yet. Other authors, however, claim that both forms have specialized semantically and that they complement each other in use, which slows down or hampers complete regularisation. One of the areas of semantic specialization is assumed to be that of aspect.

One of the first studies that looked into the effect of aspect on verb inflection is Quirk (1970), who argues that aspect might influence the selection between -ed and -t forms. More specifically, it is argued that sentences expressing duration may take the -ed form, while -t forms are said to be more frequent in punctual or effective sentences. In addition, Rohdenburg (2003: 263) states that this distinction may be iconically motivated. In other words, the ‘longer’ -ed form (both phonetically and orthographically) is associated with durativity, while -t forms are shorter and used more frequently with punctual actions.

1

Other verbs include: to bet, to broadcast, to bust, to dwell, to fit, to kneel, to knit, to lean, to light, to speed, to spill, to spoil, to strive, to sweat, to weave, to wed, to wet.

7 Quirk’s (1970) findings were based on a forced-choice selection test among British and American subjects, the results of which showed that the British subjects confirmed the preference for -t forms in punctual sentences and -ed forms in durative sentences. Levin (2009), too, found similar results for British English, but no conclusive evidence for American English, due to lack of inflectional variation in the American corpus.

This paper wants to pick up on these results and address a number of issues. First, Quirk's experimental study dates back to the seventies and has not been replicated since. In view of attested and ongoing verb regularisation processes and decreased inflectional variation between ed-/t forms (see De Clerck and Vanopstal 2015), it would be interesting to examine whether such a trend can still be attested in present-day English corpus data. While Levin's study (2009) seems to point in that direction for British English newspaper data in 2000, the findings were based on a fairly limited corpus and different results may arise from more recent data (and different registers). Secondly, to our knowledge, the impact of aspect on verb inflection has not been examined in L2 learners of English. Put differently, it still remains to be seen whether and to what extent the principle of iconicity that is said to underlie these trends is also active in the minds of L2 speakers.

To answer these questions, two types of analysis will be carried out. First of all, corpus research will show whether present-day corpus data corroborates the link between durativity and -ed forms and between punctuality and -t forms. Secondly, an elicitation experiment with first Bachelor and second Bachelor students at the Department of Applied Language Studies at Ghent University will probe into the effect of aspect on verb inflection on L2 learners.

8

2. THEORETICAL FRAMEWORK 2.1 Verb regularisation processes and inflectional variation

Lieberman (2007) described the dynamics of the evolution of the English language, focussing specifically on verb regularisation processes. Regularisation implies that irregular forms gradually disappear and evolve towards a regular form with an -ed spelling. By means of a vast quantitative study Lieberman found that out of 177 irregular verbs in Old English, only 98 remain irregular in present-day English (2007: 713). For instance, the verbs to help and to laugh were once irregular and are now solely used as regular verbs. Nevertheless, as verbs evolve towards regularity, there is a phase where the regular verb form is introduced, while the irregular form has not disappeared yet, causing both forms to co-occur. Put differently, regularisation steers verbs towards more regularity, but also allows both verb forms to co-exist for a period of time.

There is a limited set of English verbs that find themselves in this phase of the regularisation process and consequently allow both the -ed and -t verb form. For instance, the verbs to burn and to learn can be conjugated as burned/burnt and learned/learnt. Both forms, however, seem to linger on and it is not clear why these verbs have not evolved towards the regular -ed spelling, or towards the irregular -t spelling for that matter.

De Clerck & Vanopstal (2015: 347) have shown that verbs with variable inflection demonstrate different individual behaviour within different varieties of English. Some verbs appear to have a clear preference for -ed forms such as the verb to learn, while the verbs to leap and to light have a preference for -t forms. (2015: 348). Moreover, the degree of preference for each verb varies as well. The verbs to learn and to lean clearly prefer an -ed spelling with a preference of respectively 87% and 82% (2015: 348). The opposite is true for the verbs to leap and to dwell which have a strong 55% to 75% preference for -t forms (2015: 348).

9 One of the explanations for the retention of both verb forms could be that the process of regularisation has not been completed yet. However, many authors claim that both verb forms have gained a specialized meaning, which allows them to co-exist. Moreover, this semantic specialization is reportedly driven by aspect.

Kroch (1994: 1) calls this form of inflectional variation a morphological doublet. Doublets arose through dialect and language contact and compete in usage until one of them wins out. Kroch (1994: 6) observed that 'different variants can survive separately in different registers, styles or social dialects, but can only co-exist stably if they have different meanings'.

In no cases where variation can be found already in Middle English, however, do both forms survive to the present day, except where they have become differentiated in meaning. (Kroch, 1994: 7)

For instance, the morphological doublet shined and shone already existed in Old English, but differentiated in meaning and in use as one form was exclusively used as a transitive verb and the other as an intransitive one (1994: 8). As for the underlying reason, the -ed/-t doublet possibly became stable due to differentiation in meaning and grammatical properties. Levin (2009: 68) corroborates that there might be 'a functional motivation for retaining the variation' between -ed and -t forms, as the irregular form has not regularised yet. These aspects will be elaborated on in the next section.

In addition, not only individual verbs have a preference for either the -ed or -t verb form, but also the different varieties of English are biased. In his corpus study Levin (2009: 69) reached the conclusion that -t forms are favoured in British English, but seem to co-exist with -ed forms. American English, however, has reached an advanced level of regularisation, favouring the regular -ed form. These results corroborate an earlier study by Bybee and Slobin (1982: 275), which remarked that -t forms are gradually disappearing in American English.

10 Based on a comprehensive comparative study, Schlüter (2009: 110) adds that American English might be more susceptible to simplification and informality as opposed to British English and therefore features a higher number of -ed verb forms. British English is deemed more formal and as a consequence may provoke more -t verb forms (Levin, 2009: 69).

2.2 Impact of variables on -ed and -t selection 2.2.1 Aspect Aspect is a factor that reportedly influences the choice between both spelling variants and that could account for the retention of the irregular -t spelling.

One of the first studies that looked into the effect of aspect on verb inflection is Quirk (1970), who argues that aspect might influence the selection between -ed and -t forms. More specifically, it is argued that sentences expressing duration may take the -ed form, while -t forms are said to be more frequent in punctual or effective sentences, as illustrated in the examples below (Quirk 1970: 300):

(1) The gas normally smelled like bad eggs. [continually smelled like bad eggs] (2) Man may not have smelt killer gas. [effective or point-action aspect]

Quirk (1970) examined this phenomenon for the following verbs: to burn (transitive), to dream, to kneel, to lean, to leap, to learn, to smell (transitive), to spell and to spoil. His findings were based on a forced-choice selection test and a judgment test among British and American subjects. For the forced-choice selection test, the subjects were given sentence pairs, consisting of a durative and a punctual context for the same verb. They were then asked to fill in the verb form they preferred, but if they chose the -t form in the first sentence, they were obligated to fill in an ed form in the second sentence and vice versa.

11 By way of illustration the sentence pair for the verb to leap from Quirk's (1970: 302) forcedchoice test is presented below:

(3) The child...............about until it was utterly tired. (leaped, leapt) (4) The cricketer.................and managed to catch the ball. (leaped, leapt)

The results showed that the British subjects confirmed the preference for -t forms in punctual sentences and -ed forms in durative sentences. The American data, however, was less convincing due to flaws in test design. Some of the data were eliminated because some sentences were directly selected from newspaper reports and seemed unnatural when presented out of context from a posthoc perspective. For instance, the data collected for the verbs to lean and to smell were omitted from the final results. First of all, these verbs showed an opposite preference and could not 'be given the same weight as those with which the postulated distinction was observed' (1970: 306). Moreover, this opposite preference did not distinguish the American from the British subjects and therefore this data was taken out of consideration. As Quirk did not include the results for verbs with an opposite preference, his overall outcome may be too optimistic, as many researchers describe an aspectual meaning differentiation as merely a tendency. In this dissertation all results will be considered, including verbs that may not entirely support an aspectual distinction or even feature an opposite preference. Nevertheless, Quirk found the following results for the forced-choice selection test (without the results for to lean and to smell) (1970: 308) :

UK

US

-ed punctual, -t durative

97

96

-ed durative, -t punctual

245

129

Table 1 : results forced-choice selection test Quirk (1970)

12 The judgment test looked into the subjects' responses to all the possible arrangements on a classic three-point scale. The possible options were: -ed in punctual sentences; -t in durative sentences; ed in durative sentences and -t in punctual sentences. The results are shown in the table below: UK Objections to punctual -ed

US 145

16

Objections to punctual -t

39

146

Objections to durative -ed

98

22

109

156

Objections to durative -t Table 2 : results three-point scale judgment test Quirk (1970)

Given these results, Quirk (1970: 308) concluded that a general preference for -t and -ed forms in respectively punctual and durative sentences may exist for the past simple. However, further research is required with special attention to the individual behaviour of these verbs.

Levin (2009), too, paid special attention to aspect in his study. In extracts from the British newspaper the Independent, he found a significant correlation between -t forms and punctuality in the past simple for the following verbs: to burn, to leap and to learn (2009: 66). More specifically, he found that actions described by the verb to leap were mostly punctual and took a t spelling. In addition, the verb to burn occurred with an -ed spelling in 70% of the durative sentences (2009: 82). However, the opposite is shown in the case of to spill, which exhibited a significant preference for -ed forms with punctual actions. According to Levin’s corpus results the verb form spilled was used in 87% of the punctual sentences. (2009: 82) Based on Levin's results for the verb to spill, we might conclude that the verb form spilled is much more frequent and consequently more known than its irregular counterpart spilt and therefore occurs in both durative and punctual sentences (see section 2.2.4 on frequency effects p.20).

13 As for underlying results, Levin concludes (2009: 81) that aspect may have an influence on the general division of both forms, provided an individual approach is adopted. In other words, the aspectual distinction should be examined for each verb individually, as many other factors may be at stake affecting the distinction of both verb forms. For instance, in some verbs the aspectual distinction is strong, as is the case for the verb form leapt in punctual contexts, but for other verbs such as to spill the complete opposite is true. Levin also points in the direction of the different shades of meaning these verbs have. He states the following (2009: 82):

Specialization can also be investigated further with some verbs that occur with fairly distinct meanings. For instance, spell can mean 'to form by writing' (spell one's name), 'indicate something bad' (spell disaster) or 'to explain in detail' (spell it out clearly). Similarly, spoil can refer to the effects either of ruin or decay on an object or of an overindulgent upbringing on a person. We also saw some indications in section 4.1 that spill has different morphological preferences in different phrases. Such potential specializations can further support the two inflectional patterns in BrE. As mentioned in the introduction a possible aspectual distinction implies the principle of iconicity. Nänny and Fischer (2006: 462) define iconicity as ‘a sign (which may be a word or an assemblage of words), which is said to mime the object or thought that the sign refers to when something in the sign (be it oral, aural or visual) reflects something in the object that is figured by it (its referent)’. In example (5) the verb form burned can be seen as a sign that reflects the object fire and the -ed spelling can be interpreted as a feature of the sign that highlights the durative character of the object to which the sign refers.

(5) The fire (= object) in the woods burned (= sign) on. (own example) De Cuypere (2008: 104) states that the application of morphological rules results in analogy, which is merely a reproduction, and is not iconically motivated. For variant inflection to be iconically motivated both variants need to differ in meaning or have a different interpretation. In other words: ‘the observed similarity can only qualify as iconicity when the similarity adds extra meaning to the utterance or when it determines the interpretation of linguistic structure’ (2008: 104). In example (5) above, the verb form burned is a reflection of the object to which it refers, namely the fire, and the -ed spelling has an influence on the way the object is interpreted. In other words, the aspect of durativity is added to the meaning of the sentence.

14 Dingemanse et al. (2015: 603) also believe that iconicity adds meaning and define it as ‘the resemblance between aspects of form and aspects of meaning’. Based on this statement we can conclude that the morphological structure of the verb form determines its meaning. The -ed form is phonetically longer and therefore associated with durativity, while -t forms are shorter and used more frequently with punctual actions. That gives both forms a separate meaning, allowing them to co-exist. This could also be called isomorphism, which in essence is a synonym for the one form, one meaning principle Kroch refers to. On a par with Kroch, Givón (1985: 189) calls this the iconicity meta principle:

All other things being equal, a coded experience is easier to store, retrieve and communicate if the code is maximally isomorphic to the experience.

Put differently, if both forms have a clear separate meaning, they are more easily memorized and used. That implies that if the -ed spelling is generally used for durative contexts and the -t spelling for punctual contexts, both variants have a different meaning and therefore a motivated existence.

Variant inflection does not only apply to the iconicity meta principle, but also to the quantity principle. According to Givón (1991: 87) the quantity principle implies that ‘a larger chunk of information will be given a larger chunk of code’. Rohdenburg (2003: 263) adds: 'As is wellknown there is a universal tendency to code semantically more specific or more highly marked categories like the plural or the past tense by means of additional morphological devices'. As for the underlying reason, the -ed and -t form can be considered as additional code to reflect the durative or punctual character of the action the verb form refers to.

15 2.2.2 Transitivity Another important factor related to aspectual patterns is transitivity, which may also influence the choice between an -ed or -t spelling. In Levin's corpus study (2009: 68) the verb form burned was more frequently used as an intransitive verb as opposed to burnt, which was mostly transitive and combined with adverbials indicating the end of the action. According to Levin (2009: 67) intransitivity implies that there are no semantic limitations to the sentence, which produces a greater number of -ed forms. In the examples below the verb to burn is used transitively (6) and intransitively (7):

(6) I tried it once and burnt my mouth so badly I looked like I had been kissing superglue. (BNC)

(7) But the fire burned on, with the cost of the damage estimated at around sixty thousand pounds. (BNC)

Example (6) exhibits the verb to burn used intransitively and, as Levin calls it, semantically unbounded (2009: 67). In other words, there is no specific end point to the action described, highlighting its durative character. In line with the iconic principle, an -ed form is used. Example (7), on the other hand, shows the verb to burn as a transitive verb with my mouth as an object. To burn something implies a punctual or effective action which may provoke a –t spelling.

However, punctual sentences can also be intransitive (8) and durative sentences can take an object (9), such as in the examples below:

(8) Raven bent to pick the envelope up, and nudged the door shut with his hip as he unfolded it. Several pictures spilled out into his palm, and he stopped dead, feeling the air squeeze out of his lungs. (BNC)

(9) Hugh Edwards and his friends spoiled wonderful moments with their mindless wittering. (BNC)

16 Sentence (8) expresses durativity as there is an implicit indication of a habit. Hugh Edwards and his friends always spoiled wonderful moments. Wonderful moments is the object of the sentence, showing that a durative sentence does not necessarily have to be intransitive to be semantically unbounded (2009: 67). Sentence (9) expresses a punctual action: the pictures fell out of the envelope into his palm. There is no object or indication that determines the end of the action. Nevertheless, we must take into consideration that transitivity may lead to punctuality and that the absence of an object may imply a certain durativity. 2.2.3 Tense and voice Variant selection may also depend on the tense that is used. Quirk's (1970: 302) elicitation experiment showed a greater preference for irregular -t forms in the past participle, rather than the past simple. Levin's study (2009) showed an approximately equal distribution of -ed and -t forms in both British and American English. However, in one British newspaper there were significantly more -t forms in the past participle (2009: 71). Levin even went a step further by observing the variant inflection in passive sentences. He concluded that passives stimulate the use of a -t spelling in the past participle. In example (10) below a passive sentence is given which expresses a durative context. Nevertheless, a –t form is used.

(10) They wanted the lessons to be learnt and digested. (2009: 73)

These variables will not be taken into account in the data analysis. As such, we will only select simple past forms in active sentences. 2.2.4 The Principle of Rhythmic Alternation The principle of rhythmic alternation is a speech pattern that separates stressed syllables by alternating them with unstressed syllables. That way sequences of stressed (stress clashes) and sequences of unstressed syllables (stress lapses) are avoided. Lee & Gibbons (2006: 448) state that this principle 'appears to govern the distribution of strong (stressed) and weak (less stressed or unstressed) syllables in English.'

17 Lee & Gibbons (2006) have examined the phonological influence of the principle of rhythmic alternation on grammatical structures, such as the use of complementiser that. They consider the following example:

(11) Henry knew (that) Lucy washed the dishes. (12) Henry knew (that) Louise washed the dishes. (2006: 449)

Example (11) shows that the addition of complementiser that avoids the stress clash provoked by the sequence of knew Lucy. In example (12), on the other hand, the name Louise has no initial stress and therefore that is not necessary to avoid a possible stress clash.

In a recall-based spoken production experiment, native English-speaking participants were asked to omit or add that in sentences similar to the example above. They concluded that the use of complementiser that was affected by the phonological preference for rhythmic alternation in speech. In other words, in case of a stress clash, that was added to respect the phonological properties of the sentence. Schlüter (2009) examined the influence of rhythmic preferences on English verbs in particular. She mainly focussed on the verbs to light and to knit in terms of grammatical variation in British and American English. Schlüter's corpus research (2009: 113) shows that to light exhibits a tendency towards irregularisation, which is the opposite of regularisation. With regularisation, infrequent irregular forms are expected to evolve towards the more frequent regular forms. In the case of irregularisation, the irregular form is more frequent than the regular form and is therefore more likely to survive. Accordingly, the less frequent regular verb form lighted is expected to evolve to the more frequent verb form lit. Schlüter (2009: 113) suggests three categories of uses in which stress clashes or lapses could occur or are avoided: attributive uses, attributive uses with additional attributive material and non-attributive uses. First of all, the principle of rhythmic alternation is mostly applicable to attributive uses, in which stress clashes are regarded as more problematic than a stress lapse.

18 As exhibited in example (13) below the verb form lit used attributively provokes a stress clash in combination with an initially stressed noun. (13) ...; he crossed the Embankment and stood for a time watching the dark river and turning ever and again to the lít buíldings and bridges. (2009: 113)

However, the stress clash in example (13) can easily be avoided by adding additional attributive material, which is the second category Schlüter mentions. Prefixation, for instance, is a good stress clash buffer. In example (14) below prefixation is applied:

(14) We made a bolt for the únlit síde cavern forthwith. (2009: 114)

The prefix un causes the initial stress of the verb form lit to shift towards the beginning of the word. Consequently, lit becomes an unstressed syllable that alternates the stressed prefix un and the stressed noun side.

The third category implies all syntactic contexts where the verb form is used non-attributively, such as verbal uses, predicative and post-nominal uses. However, stress clashes rarely occur in this category, as they are often alternated with function words. In example (15) below, participle lit is used predicatively, but does not provoke a stress clash because of the alternation with the function word not.

(15) Had a policeman intervened because their lamps were not lít, Hoopdriver had cut him down and ridden on, after the fashion of a hero born. (2009: 114)

19 Based on the principle of rhythmic alternation and the way the verb to light acts in the previously mentioned categories, Schlüter predicted the following:

In the process of irregularization observed in Figure 5.1, lighted should disappear first where its additional syllable runs the risk of creating a stress lapse, i.e. in complex attributive structures. At the same time, it should be most tenacious where its unstressed syllable can serve as a stress clash buffer, i.e. in single unmodified attributive uses. (2009, 114) She concluded that the verb to light occurred with an -ed spelling in order to avoid a stress clash, which implies a specific reason to maintain both forms, regardless of (ir)regularisation processes. Put differently, the verb form lighted is sustained because of rhythmic alternation despite irregularisation processes favouring irregular lit. As for the underlying reasons, this alternating pattern of stressed and unstressed syllables could interfere in the choice between the monosyllabic irregular –t form and disyllabic regular –ed form in the past simple and past participle. We will examine rhythmic patterns with both the regular and irregular form of the verb to light in the examples (16) and (17) below: (16) In 1685 oil lamps already ‘lighted ´streets and parks and hackney carriages became more and more common. (BNC)

(17) Simmonds and Peacock both lít tórches from the flame in the Paralympic cauldron and marched out of the stadium, symbolically marking the end of the 2012 Games. (BNC)

In example (16) a stressed syllable is alternated with an unstressed one, which satisfies the principle of rhythmic alternation. Example (17), however, exhibits a break in the pattern by two consecutive stressed syllables, provoking a stress clash. Therefore, a stress clash could influence the choice between both spelling variants, preferring an -ed form over a -t form.

20 However, unlike the verb form lighted, irregular verb form burnt, for example, might be sustained because of aspectual differences, as there is no difference in pronunciation between the regular and irregular verb form. Nevertheless, the influence of phonological preferences on the choice between an –ed and –t spelling must be kept in mind for the sentences in the survey for the L2 learners.

2.2.5 Frequency Two types of frequency can be distinguished: token frequency and type frequency (Bybee, 1997). Type-frequency is the frequency of a pattern, for instance the use of regular –ed to form past tenses as opposed to irregular verb formation. Token-frequency is the frequency of actual items (for instance, the number of times learned occurs as opposed to learnt). This implies that the type frequency of past formation with –ed will generally be higher, as it is the default rule. However, individual verbs may still have higher token frequencies for the irregular form if both patterns are allowed. The actual token frequency of the irregular and regular forms may of course have an impact on the choices that are made: if the –ed form occurs significantly more frequently than the -t-form (e.g. 90% vs. 10%), one may assume it to occur outside contexts of durativity as well. Conversely, a high proportion of –t forms may lead to expansion from punctual contexts to durative contexts. Nevertheless, it still remains to be seen whether infrequent forms do indeed prefer specific aspectual contexts.

21

3. METHODOLOGY This dissertation will explore whether aspect is a decisive factor in the choice between the -ed and -t spelling of verbs that take both a regular and an irregular form in past conjugation. In order to answer this question, two types of analysis will be carried out. First of all, present-day corpus data will be used to corroborate whether the link between durativity and -ed forms and between punctuality and -t forms can still be attested today. Secondly, a forced-choice selection test should give an idea to what extent L2 learners of English sense this distinction. However, as the forced-choice test limits the choice of the participants, a second test was drawn up in which students were free to fill in the verb form of their preference.

The corpus-based analysis will single out British English only and involves data from the early nineties and early 21st century data. In order to control the possible impact of other variables on – ed and –t, data selection has been restricted to simple past uses in active sentences (cf. the influence on past participle and passives on the use of –ed). Finally, we only consider written text. In the BNC spoken data is included, but transcriptions do not always appear reliable and reflect the personal preferences of the transcribers.

3.1 Corpus data For the corpus-based analysis of this paper, data was drawn from the British National Corpus (BNC) and complemented with data from the Global Web-Based English corpus (GloWbE). The BNC consists of at least 100 million words from the late 20th century organised per genre, including data retrieved from newspapers, magazines, etc. GloWbE was released in 2013 and contains 1.9 billion words, retrieved from informal blogs and other website data from 20 Englishspeaking countries. The UK data consists of 387,615,074 words. The following verbs were examined: to burn, to dream, to lean, to leap, to learn, to smell, to spell, to spill and to spoil. These verbs were selected because they have all been examined before by Levin and Quirk. That way we could create a framework to compare our results to. To spill, in particular, was selected because Levin found an opposite preference, which we wanted to verify in our corpus study. In

22 addition, Quirk did not find an aspectual relation for the verb to smell and we wanted to determine whether this is still the case.

For each verb, the initial purpose was to select 100 sentences from each corpus for each verb form. However, not every verb reached the threshold of 400 sentences as there were not enough sentences at our disposal. BNC is organized in text types and not every category could be used in this study. For instance, we could not select sentences from the categories spoken and academic, as we decided to only use written data and no formal language. In GloWbE there were generally more than enough sentences we could analyze, except for the verb to spill. In table 3 below the number of sentences that were examined for every verb are listed:

To burn To learn To smell To leap To spoil To spill

BNC

GloWbE

169 200 200 200 111 145

200 200 200 200 200 168

Table 3: number of sentences that was examined in the corpus study

In addition, many sentences had to be taken out of consideration for an array of reasons. First of all, in some sentences the aspectual relation was unclear and therefore not relevant to our research. This is exemplified in the contrast between sentences (18) and (19) below:

(18) Nicholson dreamed for a day or two. (BNC) (19) She fell asleep with a smile on her face, and dreamt of Tom. (BNC)

In sentence (18) the adverbial group for a day or two clearly indicates that the sentence is durative. In sentence (19), however, there is no indication of how long the action goes on. We could argue that the verb to dream implies a certain durativity, but if we compare sentence (19) with sentence (18), it is not completely certain whether sentence (19) is also durative or not. Therefore, these sentences were not included in our corpus research. In addition, many sentences

23 were tagged wrongly (with part of speech tags for instances that were not simple past). Finally, sentences without any punctuation whatsoever were not examined either, because it was not clear where the sentence started or ended. As a consequence the data for the verbs to dream, to lean and to spell were not included in the results of this paper due to insufficient data points.

When all sentences were selected from both corpora, every sentence was labeled as durative or punctual. In many sentences the durative or punctual aspect was explicit because of an adverbial group (e.g. for days vs. suddenly). Other sentences were more implicitly durative expressing a habit or repetitive actions (When he came home from work, he smoked a cigarette in that chair). When there was no indication of habit or repetition, the sentence was interpreted as punctual. Finally, each sentence was labelled as transitive or intransitive. Interrater agreement was tested with a second annotator (B. De Clerck) and reached significantly high levels.

3.2 Survey

3.2.1 Informants For the survey-based analysis of this dissertation, a forced-choice selection test was carried out with first and second bachelor students from the Department of Applied Linguistics at Ghent University. In total there were 178 participants, 114 first Bachelor students and 64 second Bachelor students. 12 of them are Erasmus students.

3.2.2 Tests In order to answer the question if L2 learners of English sense the aspectual distinction between the -ed and -t spelling in the past simple, a forced-choice selection test was drawn up (see Quirk 1970). The informants were given sentence pairs, consisting of a durative and a punctual context for the same verb. They were then asked to fill in the verb form they preferred, but if they chose the -t form in the first sentence, they were obligated to fill in an -ed form in the second sentence and vice versa. The sentences used in the test were selected from BNC and exhibited clear aspectual properties (e.g. adverbial group, habit). Nevertheless, we added additional indications to highlight the durativity or punctuality of the sentence to avoid any doubt. The following verbs

24 were examined: to burn, to dream, to learn, to leap, to smell, to spill, to spell, to spoil, to lean and to light. For the latter verb, we tested the impact of the principle of rhythmic alternation. The informants were given a durative sentence with a stress clash, in which we expected them to use lighted and a punctual sentence without a stress clash, in which we expected the irregular verb form lit.

As the subjects' choice is limited by the forced-choice principle, another type of test was made. This second test contained 10 individual sentences, as opposed to the sentence pairs in the forcedchoice, in which the subjects could freely choose between the -ed form and -t form (from here on referred to as free-choice test). Of this free-choice test two versions were made. For instance, a durative sentence for the verb to burn was provided in version 1 and a punctual sentence in version 2. If we provided both a durative and a punctual sentence in the same version, students might have filled in the same verb form twice by way of analogy. However, in the case of the verb to light, we provided the same punctual sentence in both versions to probe into the impact of a stress clash on students' choice between the -ed and -t spelling. The sentences used in this test are the same as in the forced-choice, but were divided over the two separate tests. The punctual and durative contexts were divided over both tests as followed:

Version 1

Version 2

durative burn

punctual burn

punctual learn

durative learn

durative smell

punctual smell

punctual leap

durative leap

durative spell

punctual spell

punctual dream

durative dream

durative spill

punctual spill

punctual spoil

durative spoil

durative lean

punctual lean

punctual light

punctual light

Table 4: division punctual and durative sentences

25 In the test several variables had be kept stable. First of all, the durative and punctual sentences were alternated to avoid patterns. Secondly, both verb forms were given in the instructions and in the sentences of the test to guarantee that students knew that both verb forms are equally correct in every sentence. Finally, all sentences of both tests were altered so as to satisfy the principle of rhythmic alternation. In other words, all stress clashes were avoided to make sure that this principle does not provoke results that might give the wrong idea of the impact of aspect. However, in the case of to light a punctual sentence with a stress clash is given in every test, in order to examine whether students fill in an -ed form to respect the rhythmic properties of the sentence or a -t form to respect punctuality.

26

4. RESEARCH QUESTIONS Due to the experimental nature of this paper it is hard to predict the outcome of both the corpus and survey data. Moreover, many researchers describe this phenomenon as merely a tendency. Therefore, instead of formulating hypotheses, we will limit ourselves to answering the following research questions:

4.1 Corpus data  Do sentences with –ed have a tendency of occurring with durative aspect?  Do sentences with –t have a tendency of occurring with punctual aspect?  Does the verb’s general preference for -ed/-t have an impact on the strength of the tendency?  Are there differences between the BNC and GloWbE data?

4.2 Survey data  Does the principle of iconicity manifest itself in L2 learners through a correlation between -ed and durativity and -t and punctuality?  Does a stress clash have an influence on the choice between an -ed or a -t verb form? Put differently, what is deemed more important? Aspect or complying with the principle of rhythmic alternation?  Do the results of the forced-choice test differ from those of the free-choice test?  How can divergence from the iconicity principle be explained?

27

5. DATA ANALYSIS 5.1 Corpus data 5.1.1 General results Table 4 presents the general results from the corpus data.

-ed Durative

-t Punctual

Durative

Punctual

BNC

260

46%

302

54%

172

37%

291

63%

GloWbE

213

36%

387

64%

216

38%

352

62%

Total

473

41%

689

59%

388

38%

643

62%

Table 5: general results corpus study

The general overview shows that there is no clear discrepancy between –t forms and –ed forms regarding aspectual preferences. In both cases, roughly 60% occurs with punctual uses, while durative uses account for 40% of the data. A Fisher’s exact test revealed no statistical difference (p = 0.14). In the BNC results, however, the division of durativity and punctuality in -ed sentences seems to be more equally divided. This equal division disappears in the more recent GloWbE data in which approximately 10% more sentences were labelled punctual.

These results do not immediately point in the direction of an aspectual difference between both verb forms. However, it would be wrong to generalize these findings, as we have established in the literature study that individual verbs behave differently and may have different preferences (See De Clerck and Vanopstal 2015). Therefore, we will discuss the attested preferences for each individual verb that was examined in this corpus study and put these general results into perspective (see appendix 4).

28 5.1.2 Results individual verbs The BNC and GloWbE corpus results are shown in Table 6 below. The verbs are organized in accordance with the highest number of either punctual sentences with -t verb forms or durative sentences with -ed verb forms. The BNC data is presented first, followed by the more recent GloWbE data in both absolute figures and proportions. In addition, Table 7 shows the attested proportions of -ed and -t verb forms in GloWbE (De Clerck & Vanopstal, 2015: 348). Figures confirming the link between aspect and verb form are marked in bold.

-ed Durative To spill To leap To burn To smell To spoil To learn

46 25 18 14 49 53 73 75 15 31 59 15

46% 25% 18% 14% 49% 53% 73% 75% 24% 31% 59% 15%

-t Punctual 54 75 82 86 51 47 27 25 47 69 41 85

54% 75% 82% 86% 51% 47% 27% 25% 76% 69% 41% 85%

Durative 5 5 8 12 9 15 62 71 18 33 70 80

11% 7% 8% 12% 13% 15% 62% 71% 37% 33% 70% 80%

Total Punctual 40 63 92 88 60 85 38 29 31 67 30 20

89% 93% 92% 88% 87% 85% 38% 29% 63% 67% 30% 20%

Durative

Punctual

51 30 26 26 58 68 135 146 33 64 129 95

94 138 174 174 111 132 65 54 78 136 71 105

Table 6: corpus results individual verbs

-ed To burn To leap To learn To smell To spill To spoil

11,675 759 67,035 1,803 2,078 3111

-t 70% 32% 82% 72% 79% 68%

5,081 1,580 15,212 715 554 1434

30% 68% 18% 28% 21% 32%

Table 7: attested proportions -ed/-t verb forms GloWbE (De Clerck & Vanopstal)

35% 18% 13% 13% 34% 34% 68% 73% 30% 32% 65% 48%

65% 82% 87% 87% 66% 66% 32% 27% 70% 68% 35% 52%

29 For the verb to leap the BNC and GloWbE results appear to be unanimous. 80-90% of all sentences, in which either leaped or leapt was used, were labelled punctual. This might prove that native speakers have a clear preference for the irregular verb form leapt in punctual sentences, confirming our hypothesis. However, it does not explain why the -ed verb form also features in a majority of punctual sentences instead of durative ones. Other factors may be at play, one of which might be the verb’s meaning. The Oxford Dictionary describes to leap as 'to move suddenly or quickly' or 'to jump across', which are typically actions one would associate with punctuality. As such, the meaning of the verb itself carries punctuality and leads to a high concentration of punctual uses, regardless of the actual verb form.

The same applies to the results of the verb to spoil. Out of all sentences with irregular verb form spoilt, 60-70% were punctual, confirming natives' preference for -t verb forms in punctual sentences. However, regular verb form spoiled shows an even higher number of punctual sentences, namely 70-80%. These results may be explained by the fact that the -ed form in itself is a lot more frequent than the -t form (see GloWbE data: 68% vs. 32%), which could also imply that the regularized form will also be attested more frequently in punctual contexts.

In accordance with the results for to leap and to spoil, there are also more punctual contexts for -t verb forms in the case of to spill and to burn. However, unlike the results for to leap and to spoil, the number of durative and punctual sentences with -ed verb forms seems to be more equally divided in the BNC data. This may confirm our hypothesis that -t verb forms are preferred in punctual sentences, while -ed verb forms are preferred in durative sentences. The preference for ed verb forms might not reach 50%, but the retention of this fraction might point in the direction of an aspectual meaning differentiation. The GloWbE data for to burn even reaches a slight preference for -ed verb forms in durative sentences and a strong preference for -t verb forms in punctual sentences, confirming our hypothesis. These results confirm a trend in which the most frequent form is attested in both contexts (in this case burned which accounts for 70% of the GloWbE data), whereas aspectual specialization is attested with the least frequent form: -t clearly shows a preference for punctual contexts.

30 For sentences in which the regular verb form smelled was used, 70-80% of those were labeled durative, confirming our hypothesis. However, out of all sentences with irregular verb form smelt, despite its low frequency in GloWbE, 60-70% were also labelled durative. In other words, both verb forms are used more in durative contexts than in punctual ones. The Oxford Dictionary distinguishes two main meanings for this verb: 'to perceive or detect the odour or scent of', which is transitive and 'to emit an odour or a scent', which is intransitive. One might interpret the first meaning as a punctual action and the second as a more permanent state. The sentences in our corpus were labeled accordingly. Out of the 400 sentences for both verb forms 270 were intransitive and consequently labeled as durative. We might conclude that the verb to smell is used more often in the durative meaning of the word.

In the BNC approximately 60% of the sentences with verb form learned are durative, an amount which is even 10% higher in sentences with irregular verb form learnt. However, the opposite is true for the more recent GloWbE data, which shows an opposite tendency. Nevertheless, given the attested proportions for -ed and -t verb forms in GloWbE (82% and 18%), it is noteworthy that, despite its low frequency within the corpus, irregular verb form learnt features a majority of durative sentences. This is not a case of semantic specialization, but it shows that the verb to learn is more frequent as a durative verb, namely to learn as a durative process rather than a punctual discovery.

Summing up, to burn and to spill confirm a trend of semantic specialization. The most frequent ed verb form occurs in both durative and punctual sentences, while the low-frequent -t verb form is mainly used in punctual sentences. The other four verbs in this corpus study show either a majority of durative or punctual contexts. To leap is mostly used in punctual contexts, as it bears punctuality in its meaning. The same is applicable to the verb to spoil, of which the -ed verb form is the most frequent and therefore occurs in both durative and punctual sentences. The opposite is true for to learn and to smell which feature a majority of durative contexts, which might prove that these verbs are mainly used in the durative meaning of the word.

31 5.2 Survey data

5.2.1 Forced-choice test The main aim of the survey-based part of this paper was to verify whether L2 learners of English also sense a possible aspectual distinction between the -ed verb form and durativity and the -t verb form and punctuality. To achieve this goal, two kinds of surveys were drawn up: a forcedchoice selection test and a test with individual sentences in which students could freely choose between both verb forms, the free-choice test.

In the forced-choice selection test students had to fill in both verb forms in every sentence pair. If they chose an -ed form in the first sentence, they were obligated to fill in a -t form in the second sentence and vice versa. This test was only carried out among first Bachelor students. In general, 63% of the sentences reflected the iconicity principle, which means they opted for an -ed verb form in the durative sentence and a -t verb form in the punctual sentence. In 37% of the sentence pairs students chose for a -t verb form in the durative sentence and an -ed verb form in the punctual sentence. Table 8 below shows the individual results for every verb presented in the forced-choice test:

-ed durative, -t punctual To light To learn To lean To dream To spill To spell To spoil To leap To burn To smell

35 51 49 48 42 41 39 36 35 34

79% 75% 72% 71% 62% 60% 57% 53% 51% 50%

Table 8: results forced-choice test Bachelor 1

-t durative, -ed punctual 33 17 19 20 26 27 29 32 33 34

21% 25% 28% 29% 38% 40% 43% 47% 49% 50%

32 Given these results, it appears that every verb confirms the hypothesis that -ed forms are preferred in durative sentences and -t forms in punctual ones. In other words, if we obligate students to make a choice and to use both verb forms, the results point in the direction of an aspectual meaning differentiation. However, it is not clear why there are so many internal differences between these results. The proportions of -ed and -t verb forms in durative and punctual sentences for the verb to light, for instance, differ considerably from those of the verb to smell, which shows an equal division. In the next section we will have a look at the results of the free-choice test.

5.2.2 Free-choice test The general outcome of this second test is presented in Table 9 below. However, it should be mentioned that for the verb to light, of which we wanted to test the impact of a stress clash and not necessarily that of aspect, all students, except 4, picked the irregular -t verb form. As there was no durative sentence provided for this verb in the test, it might not be completely accurate to include these results in the general outcome. Consequently, Table 9 does not include the results for the verb to light.

Durative

Punctual

-ed 347

70%

-t 152

-ed 30%

295

60%

-t 193

40%

Table 9: general results free-choice test Bachelor 1 and 2

In the durative sentences students have a 70% preference for -ed verb forms and merely a 30% preference for -t verb forms, confirming our hypothesis. In the punctual sentences, there is a 60% preference for the regular -ed verb form.

From the general outcome, we may conclude that students have an overall preference for -ed verb forms in durative sentences, but also in punctual ones. This is in line with attested regularisation processes and with our expectation that the higher type frequency of the regular -ed conjugation might cause students to apply the general rule more easily. Nevertheless, we may not jump to

33 conclusions on the basis of these general results. First we need to take a closer look at the choices students have made for the individual verbs that appeared in the test.

5.2.3 Results individual verbs Based on the results shown in Table 10 below, none of the verbs in the test features a distinctive preference for -ed forms in durative sentences and -t forms in punctual sentences. Students appear to have either a preference for the -ed or -t verb form regardless of aspect. Students (both Bachelor 1 and 2) show a general preference for the -ed verb form with the following verbs: to lean, to learn, to smell, to spell, to spill and to spoil. Students have an overall preference for -t verb forms with to dream, to leap and to light. The results for these verbs are presented in Table 10 below (for more detailed results see appendix 5 on page 50). Table 11 on page 34 shows the attested proportions of -ed and -t verb forms in GloWbE (De Clerck & Vanopstal, 2015: 348).

B

Durative -ed

Punctual -t

-ed

-t

To smell (79% -ed/ 21% -t)

52

87%

8

13%

36

71%

15

29%

To lean (69% -ed/ 31% -t)

49

82%

11

18%

27

54%

23

46%

To spell (81% -ed/ 19% -t)

49

82%

11

18%

40

80%

10

20%

To spoil (79% -ed/ 21% -t)

41

82%

9

18%

46

77%

14

13%

To learn (68% -ed/ 32% -t)

35

70%

15

30%

40

67%

20

33%

To spill (72% -ed/ 28% -t)

42

70%

18

30%

36

73%

13

27%

To burn (57% -ed/ 43% -t)

32

54%

27

46%

30

60%

20

40%

To leap (38% -ed/ 62% -t)

24

48%

26

52%

18

30%

42

70%

To dream (41% -ed/ 59% -t)

23

46%

27

54%

22

37%

38

63%

4

4%

106

96%

To light (4% -ed/ 96% -t)

-

-

Table 10: individual results free-choice test Bachelor 1 and 2 combined

34 -ed To burn To dream To lean To leap To learn To light To smell To spell To spill To spoil

11,675 5,250 3,320 759 67,035 523 1,803 3,414 2,078 3111

-t 70% 72% 87% 32% 82% 6% 72% 73% 79% 68%

5,081 2,025 508 1,580 15,212 7,537 715 1,256 554 1434

30% 28% 13% 68% 18% 94% 28% 27% 21% 32%

Table 11: attested proportions -ed/-t verb forms GloWbE (De Clerck & Vanopstal)

Students appear to have a preference for the -ed verb form in both durative and punctual sentences for the following verbs: to learn, to spell and to spill. To learn and to spill have a 70% preference for -ed verb forms in both durative and punctual sentences. For the verb to spell students have an even stronger preference (80%) in both durative and punctual contexts. Given these results, we may conclude that the -ed verb form for these verbs is more frequent and generally more known among students, which is in line with attested regularisation processes. Moreover, if the -ed verb form is the most frequent, it is highly likely that it will be used in both durative and punctual contexts. This is also corroborated by the attested proportion for -ed forms in GloWbE which reaches approximately 70-80% for these three verbs.

For the verbs to lean, to smell and to spoil students also have a preference for the -ed verb form in both durative and punctual contexts. This is in line with the attested proportions in GloWbE, which shows an approximate 70-80% for -ed verb forms. As a consequence, it may be logical that the regular verb form occurs in both durative and punctual contexts. However, in the punctual sentences the remaining fraction of -t verb forms appears to be higher than the remaining -t verb forms in durative sentences. For instance, for to lean students have an 82% preference for -ed verb forms and a 12% preference for -t verb forms in durative sentences. In punctual sentences, the preference for -ed verb forms is 54% and there is a 46% preference for -t verb forms. As there is a higher remaining fraction of -t verb forms in punctual sentences than in durative ones (respectively 46% vs. 12%), we might conclude that a number of students filled in an -ed verb form in the durative sentence and a -t verb form in the punctual one. So even though

35 they prefer the -ed verb form, these results might still show that a large group of students senses the aspectual difference between both verb forms. The same phenomenon could be attested for to spell and to spoil, though be it to a lesser extent.

Students prefer to use a -t verb form for the verbs to dream and to leap. They used -ed and -t verb forms respectively 40% and 60% of the time for both verbs. Students appear to have a 70% preference for leapt in punctual sentences. However, the remaining fraction of -ed verb forms is higher in the durative sentences than in the punctual ones (respectively 48% vs. 30%). So even though there is a preference for -t verb forms in durative sentences, approximately half of all participants still opts for an -ed verb form in durative sentences, which may indicate that they might comply with an aspectual distinction. The same is applicable to the verb to dream.

Nevertheless, it should be noted that this convinced preference for -t verb forms may be a display of hypercorrective behaviour. The salience of change from the irregular form with a short vowel to the regular form with a longer vowel/different pronunciation might make them more aware of the regularised form, which makes the irregular the safest option in the learning environment the tests were conducted in. .

For the verb to burn students filled in an -ed verb form and -t verb form in respectively 57% and 43% of the sentences, regardless of aspect. They appear to have a preference for the -ed verb form in both durative and punctual sentences. However, their -ed preference is more convinced in punctual sentences, while we can only attest a slight preference in durative sentences. The remaining fraction of -t verb forms is also higher in durative sentences than in punctual ones. We might argue that the presence of both verb forms in both durative and punctual sentences is due to the fact that burned and burnt, bear little differences in orthography and pronunciation, which does not provoke hypercorrective behaviour. The opposite is true for the verbs to dream and to leap (see previous section). This may explain why there is a more equal division of the verb forms burned and burnt.

36 The main motivation for including the verb to light, was to observe the impact of a stress clash on the students' choice between -ed and -t verb forms. In both version 1 and 2 of the free-choice test a punctual sentence was given with a stress clash. In order to satisfy the principle of rhythmic alternation, students were expected to fill in an -ed verb form. However, almost every participant chose the irregular verb form lit, even though this choice maintained the stress clash. Only 4 first Bachelor students chose for the regular verb form lighted and complied with the principle of rhythmic alternation. There may be two reasons for this. First, one could assume aspect to outweigh a stress clash as a variable. Secondly, and perhaps more importantly, the strong preference for -t verb forms may also be a display of hypercorrection. Just like to leap and to dream, both verb forms of to light are very different in spelling and pronunciation. Within the framework of their education, students have studied the irregular verbs and are very aware of the irregular spelling in the past. Consequently, students may have chosen for the irregular lit, because they are certain this verb form is correct.

37 5.2.4 Comparison Bachelor 1 and Bachelor 2 Table 12 shows the general outcome of the free-choice test, taking aspect into consideration. Table 13 provides a comparison between the data from the first and second Bachelor students, based on the free-choice test. The results for the first Bachelor students are presented first, followed by the results of the second Bachelor students.

Durative

Punctual

-ed 141 206

-t 70% 69%

61 91

-ed 30% 31%

122 173

-t 58% 62%

89 106

42% 38%

Table 12: general outcome free-choice test Bachelor 1 and 2

Durative -ed To smell To lean To spell To spoil To learn To spill To burn To leap To dream To light

18 34 17 32 17 32 21 20 18 17 14 28 11 21 12 12 13 10

95% 83% 90% 78% 89% 78% 78% 87% 67% 74% 74% 68% 61% 51% 44% 52% 48% 44% -

Punctual -t

1 7 2 9 2 9 6 3 9 6 5 13 7 20 15 11 14 13

-ed 5% 17% 10% 22% 11% 22% 22% 13% 33% 26% 26% 32% 39% 49% 56% 48% 52% 56%

-

Table 13: results free-choice test Bachelor 1 and 2

19 17 16 11 19 21 15 31 13 27 15 21 14 16 4 14 7 15 4 0

68% 74% 59% 48% 70% 91% 79% 76% 68% 66% 58% 91% 52% 70% 21% 34% 37% 37% 9% 0%

-t 9 6 11 12 8 2 4 10 6 14 11 2 13 7 15 27 12 26 42 64

32% 26% 41% 52% 30% 9% 21% 24% 32% 34% 42% 9% 48% 30% 79% 66% 63% 63% 91% 100%

38 The general outcome of the free-choice test appears to be unanimous in both Bachelor 1 (BA1) and Bachelor 2 (BA2). In durative sentences both BA1 and BA2 have a 70% preference for -ed verb forms. In punctual sentences this -ed preference remains but in smaller proportions. However, it is noteworthy that the fraction of -t verb forms in punctual sentences is greater than in durative sentences. This means that the most frequent verb form is used in both contexts, while the low frequent -t verb form is mainly used in punctual sentences. This general outcome might confirm a semantic differentiation between both verb forms that is driven by aspect. Nevertheless, we may not generalize these results without comparing the results of the individual verbs.

There are six verbs in the test that prove that the most frequent verb form is used in both contexts, while the low-frequent verb form is mainly used in one (in both BA1 and BA2). These verbs are: to smell, to lean, to spell, to spill, to leap and to dream. In the case of to lean, for instance, BA1 shows a preference for -ed verb forms, which are used in both durative and punctual contexts, while the low-frequent -t verb form is mainly used in punctual contexts. For to lean BA2 shows the same tendency to the extent that it even confirms the hypothesis that -ed verb forms are preferred in durative sentences and -t verb forms in punctual sentences. The same is applicable to the verbs to spill and to spell, with the only difference that BA2 shows an even greater -ed preference in punctual sentences than in durative ones. To leap and to dream also show the same trend, but here the irregular -t verb form is the most frequent.

As opposed to the verbs mentioned above, both BA1 and BA2 show a clear -ed preference for to spoil and to learn. In this case the low-frequent fraction of -t verb forms is the same in durative and punctual sentences.

For the verb to light, only four BA1 students chose for lighted, while the rest picked irregular lit. In BA2, every student chose the irregular -t verb form. That is probably the case because irregular lit is far more known among students than its regular counterpart lighted. So even though it was mentioned that both verb forms were correct, students might still see the irregular verb form as a safe option, which is a sign of hypercorrective behaviour.

39 In sum, it is noteworthy that BA2 students generally use more -t verb forms than BA1 students. As they have a year more of experience writing and translating texts into British English, they might sense that the -ed verb form is more American and therefore more colloquial, while the -t verb form is deemed more formal. However, this theory does not completely cover participants' behaviour, as many BA2 students use more -t in durative sentences, but also more -ed verb forms in punctual sentences for four verbs in the test (to burn, to spell, to spill and to smell). Nevertheless, where BA1 students tend to have either an -ed preference or a -t preference for a specific verb, BA2 students have it too, but in more nuanced proportions.

40

6. CONCLUSION In this paper aspect was examined as a form of semantic specialization that retains low-frequent fractions of -ed and -t verb forms of verbs with variant inflection. Two types of analysis were carried out. Firstly, corpus research had to verify whether present-day corpus data corroborates the link between -t verb forms and punctuality and between -ed verb forms and durativity. Secondly, a forced-choice and a free-choice test were carried out among first Bachelor and second Bachelor students at the Department of Applied Language Studies in order to probe into the effect of aspect on verb inflection on L2 learners.

The BNC and GloWbE data have shown a possible aspectual link between -t verb forms and punctuality, however, the exact same tendency could also be attested for -ed verb forms. Then we took a closer look at the behaviour of the individual verbs in our corpus study. The verbs to leap and to spoil feature a majority of punctual sentences for both -t and -ed verb forms. To spill and to burn might actually confirm a semantic specialization that is driven by aspect as the most frequent -ed verb form is used in both durative and punctual contexts, while the low-frequent -t verb form is mainly used in punctual contexts. As opposed to the verbs to spill and to burn, there is a higher number of durative sentences with both verb forms in the case of to smell and to learn.

The survey-based part of this paper consists of a forced-choice test and a free-choice test among first Bachelor and second Bachelor students. Every verb that appeared in the forced-choice test confirmed our hypothesis that both verb forms may bear different aspectual connotations and that students sense this distinction. So, if we obligated them to make a choice, they preferred an -ed verb form in the durative sentence and a -t verb form in the punctual sentence. However, these forced-choice results showed a high level of internal variation. These differences are confirmed in the free-choice test, in which students made different choices.

The general outcome of the free-choice test showed that students have a preference for the -ed verb form in both punctual and durative contexts. However, if we looked into the individual results for every verb, it seemed that students appear to have a preference for either an -ed or -t verb form. Students have an overall preference for the regular -ed verb form with the following

41 verbs: to lean, to learn, to smell, to spell, to spill and to spoil. These results are in line with attested regularisation processes, which makes it highly likely that -ed verb forms will occur in both durative and punctual sentences. With the verbs to dream, to leap and to light, students prefer the irregular -t verb form. The latter verb was included in the test to examine the influence of a stress clash on the choice between both verb forms and not necessarily that of aspect. From their convinced choice for irregular verb form lit we may conclude that the impact of aspect outweighs that of the principle of rhythmic alternation.

However, there are four verbs of which the remaining fraction of either -ed or -t verb forms does point in the direction of a semantic specialization, namely to lean, to smell, to leap and to dream. For instance, to lean has a convinced preference for -ed verb forms in durative sentences, while there is a more balanced division between both verb forms in punctual sentences. That might show that a large group of students did comply with the principle of iconicity.

In addition, it is noteworthy that second Bachelor students generally use more -t verb forms than first Bachelor students. As they have a year more of experience writing and translating texts into British English, they might sense that the -ed verb form is more American and therefore more colloquial, while the -t verb form is deemed more formal. Nevertheless, it could also be a display of hypercorrective behaviour. As they are very aware of irregular verb forms that differ considerably from its regular counterpart, they might choose this form because they are certain it is correct.

In sum, there are signs that an aspectual meaning differentiation between both verb forms might be possible, but there are many other factors that may account for this choice (internal variation, meaning of the verb, frequency). Nevertheless, this type of research offers possibilities for the future. It might be interesting to conduct a forced-choice and a free-choice selection test among native speakers, for instance. In addition, it would be interesting to look into the different meanings of the verbs and their link with either durativity and punctuality.

42

7. BIBLIOGRAPHY Bybee, J. L., & Slobin, D. I.. (1982). Rules and Schemas in the Development and Use of the English past Tense. Language, 58(2), 265–289. http://doi.org/10.2307/414099 Bybee, J. & Thompson, S. (1997). Three Frequency Effects in Syntax. Annual Meeting of the Berkeley Linguistics Society,23 (1), 378-388. doi: http://dx.doi.org/10.3765/bls.v23i1.1293 De Clerck, B., & Vanopstal, K. (2015). Patterns of regularisation in British, American and Indian English: a closer look at irregular verbs with t/ed variation. In P. Collins (Ed.), Grammatical change in English world-wide (pp. 335–372). Amsterdam, The Netherlands: Benjamins. De Cuypere, L. (2008). Limiting the iconic: from the metatheoretical foundations to the creative possibilities of iconicity in language. Iconicity in Language and Literature (Vol. 6). Amsterdam, The Netherlands: John Benjamins. Dingemanse, M., Blasi, D.E., Lupyan, G., Christiansen, M.H.& Monaghan, P. (2015). Arbitrariness, iconicity and systematicity in language. Trends in Cognitive Sciences, 19, 10, 603-615 Gibbons, J. & Lee, M.W. (2007). Rhythmic alternation and the optional complementiser in English: New evidence of phonological influence on grammatical encoding. Cognition, 105 (2), 446-456. Givón, T. (1985). Iconicity, isomorphism and non-arbitrary coding in syntax. In Haiman J. (Ed), Iconicity in syntax. Amsterdam: Benjamins. Kroch, A. (1994). Morphosyntactic Variation. In K. Beals et al. (Eds), Papers from the 30th Regional Meeting of the Chicago Linguistics Society: Parasession on Variation and Linguistic Theory. Chicago: Chicago Linguistic Society. Levin, M. (2009). The formation of the preterite and the past participle. In G. Röhdenburg & J. Schlüter (Eds.), One Language, Two Grammars (60-85). Cambridge: University Press Cambridge

43 Lieberman et al.(2007). Quantifying the evolutionary dynamics of language. Nature, 499. Retrieved from http://www.nature.com/nature/journal/v449/n7163/full/ nature06137.html Nänny, M. & Fischer, O. (2006). Iconicity: Literary texts? In Encyclopedia of Language & Linguistics, Vol 5, 2nd edn, K. Brown (ed. in chief), 462-472. Oxford: Elsevier. DOI: 10.1016/B0-08-044854-2/00516-2 Quirk, R. (1970). Aspect and Variant Inflection in English Verbs. Language, 46(2), 300– 311. http://doi.org/10.2307/412280 Rohdenburg, G. (2003). Aspects of grammatical iconicity in English. In W. G. Müller & O. Fischer (Eds.), From sign to signing (263-285). Amsterdam: John Benjamins Publishing. Schlüter, J. (2009). The Phonology and grammar. In G. Röhdenburg & J. Schlüter (Eds.), One Language, Two Grammars (108-129). Cambridge: University Press Cambridge To leap (n.d.). In The Oxford Dictionaries. Retrieved from http://www.oxforddictionaries.com/definition/english/leap?q=to+leap To smell (n.d.). In The Oxford Dictionaries. Retrieved from http://www.oxforddictionaries.com/definition/english/smell?q=to+smell

44 Appendix 1: version 1 of the free-choice test

In British English, some verbs can have both an –ed and a –t form in the Simple Past and the Past Participle. Put a circle around the verb form you prefer in the contexts below. Please pick one form only. Thank you for your cooperation!

1. Martin suddenly learned - learnt that another team had beaten him to it. 2. So you dreamed - dreamt up a message on the spot? You must find our evenings together remarkably predictable. 3. But Dot knew he was a doctor because his hands smelled - smelt clean and soapy. 4. But the fire burned - burnt on, with the cost of the damage estimated at around sixty thousand pounds. 5. She leaned - leant against the car for some time, which made it easier on the attacker to grab her and pull her into the vehicle. 6. Two joyriders leaped - leapt out of the moving car causing it to smash into a house. 7. Simmonds and Peacock both lighted - lit torches from the flame in the Paralympic cauldron and marched out of the stadium, symbolically marking the end of the 2012 Games. 8. Don't ignore the fact that although Michael misspelled 14 words, he consecutively spelled - spelt 75 correctly, and that his story is worth reading. 9. The water spilled - spilt out all day until the ceiling gave way. 10. Anne would hate it if she spoiled - spoilt the surprise for her sister's birthday.

45 Appendix 2: version 2 of the free-choice test

In British English, some verbs can have both an –ed and a –t form in the Simple Past and the Past Participle. Put a circle around the verb form you prefer in the contexts below. Please pick one form only. Thank you for your cooperation!

1. For days flames leaped - leapt 10 feet from windows and around 15 to 20 feet from the roof. 2. Maggie leaned - leant the broom against the wall and picked up the phone. For some reason she knew it was an emergency. 3. For a second, she smelled - smelt it: the sweet scent of the gillyflowers. And then it was gone again. 4. I learned - learnt so much in France. They were the happiest four years of my life. 5. I tried it once and burned - burnt my mouth so badly I looked like I'd been kissing superglue. 6. For decades, the villagers dreamed - dreamt about having a well of their own, but it never happened. 7. Amazingly, Sophie spelled - spelt the word correctly and without a moment's hesitation. 8. When I shouted he spilled - spilt his coffee. 9. Dad spoiled - spoilt his grandkids always giving them lottery tickets for Christmas hoping at least one could get back as much as he had wasted. 10. Simmonds and Peacock both lighted - lit torches from the flame in the Paralympic cauldron and marched out of the stadium, symbolically marking the end of the 2012 Games.

46 Appendix 3: Forced-choice test In British English, some verbs can have both an –ed and a –t form in the Simple Past and the Past Participle. Put a circle around the verb form you prefer in the contexts below. If you choose the -t form in the first sentence, you have to choose the ed-form in the second sentence and vice versa. Make sure to read both sentences first before making a decision. Thank you for your cooperation!

1. But the fire burned - burnt on, with the cost of the damage estimated at around sixty thousand pounds. I tried it once and burned - burnt my mouth so badly I looked like I'd been kissing superglue.

2. So you dreamed - dreamt up a message on the spot? You must find our evenings together remarkably predictable. For decades, the villagers dreamed - dreamt about having a well of their own, but it never happened.

3. Martin suddenly learned - learnt that another team had beaten him to it. I learned - learnt so much in France. They were the happiest four years of my life.

4. She leaned - leant against the car for some time, which made it easier on the attacker to grab her and pull her into the vehicle. Maggie leaned - leant the broom against the wall and picked up the phone. For some reason she knew it was an emergency.

5. But Dot knew he was a doctor because his hands smelled - smelt clean and soapy. For a second, she smelled - smelt it: the sweet scent of the gillyflowers. And then it was gone again.

47 6. Two joyriders leaped - leapt out of the moving car causing it to smash into a house. For days flames leaped - leapt 10 feet from windows and around 15 to 20 feet from the roof.

7. Don't ignore the fact that although Michael misspelled 14 words, he consecutively spelled - spelt 75 correctly, and that his story is worth reading. Amazingly, Sophie spelled - spelt the word correctly and without a moment's hesitation.

8. Anne would hate it if she spoiled - spoilt the surprise for her sister's birthday. Dad spoiled - spoilt his grandkids always giving them lottery tickets for Christmas hoping at least one could get back as much as he had wasted.

9. The water spilled - spilt out all day until the ceiling gave way. When I shouted he spilled - spilt his coffee.

10. Aunty lighted - lit a cigarette, took a deep drag, and began to speak, sadness creeping into her voice. 'Marriages,' she said, 'are made in Heaven'. In 1685 oil lamps already lighted - lit streets and parks and hackney carriages became more and more common.

48 Appendix 4 : results corpus study BNC GloWbE Total

To burn

To leap

To learn

To smell

To spill

To spoil

All

-ed durative 49 53 102 18 14 32 59 15 74 73 75 148 46 25 71 15 31 46

49% 53% 51% 18% 14% 16% 59% 15% 37% 73% 75% 74% 46% 25% 36% 24% 31% 28%

172 216 388

37% 38% 38%

-t durative

9 15 24 8 12 20 70 80 150 62 71 133 5 5 10 18 33 51 291 352 643

13% 15% 14% 8% 12% 10% 70% 80% 75% 62% 71% 67% 11% 7% 9% 37% 33% 34% 63% 62% 62%

punctual 60 85 145 92 88 180 30 20 50 38 29 67 40 63 103 31 67 98 260 213 473

87% 85% 86% 92% 88% 90% 30% 20% 25% 38% 29% 33% 89% 93% 91% 63% 67% 66% 46% 36% 41%

durative 49 53 102 18 14 32 59 15 74 73 75 148 46 25 71 15 31 46 302 387 689

49% 53% 51% 18% 14% 16% 59% 15% 37% 73% 75% 74% 46% 25% 36% 24% 31% 28% 54% 64% 59%

Appendix 5: results free-choice and forced-choice test

Ba1: version 1+2+FC Ba2: version 1+2 Total to burn ED 57% /T 43% To dream ED 41% /T 59% To lean ED 69% /T 31% To leap ED 38% /T 62% To learn ED 68% /T 32%

Durative -ed 11 21 32 13 10 23 17 32 49 12 12 24 18 17 35

To light

To smell ED 79%/ T 21%

61% 51% 54% 48% 44% 46% 90% 78% 82% 44% 52% 48% 67% 74% 70%

-t 7 20 27 14 13 27 2 9 11 15 11 26 9 6 15

18 34 52

Punctual -ed 39% 49% 46% 52% 56% 54% 10% 22% 18% 56% 48% 52% 33% 26% 30% -

95% 83% 87%

1 7 8

5% 17% 13%

14 16 30 7 15 22 16 11 27 4 14 18 13 27 40 4 0 4 19 17 36

52% 70% 60% 37% 37% 37% 59% 48% 54% 21% 34% 30% 68% 66% 67% 9% 0% 4% 68% 74% 71%

Forced-choice -t

13 48% 7 30% 20 40% 12 63% 26 63% 38 63% 11 41% 12 52% 23 46% 15 79% 27 66% 42 70% 6 32% 14 34% 20 33% 42 91% 64 100% 106 96% 9 32% 6 26% 15 29%

Yes

No

35

51%

33

49%

48

71%

20

29%

49

72%

19

28%

36

53%

32

47%

51

75%

17

25%

54

79%

14

21%

34

50%

34

50%

50

To spell ED 81%/ T 19% To Spill ED 72%/ T 28% To spoil ED 79%/ T 21%

17 32 49 14 28 42 21 20 41

89% 78% 82% 74% 68% 70% 78% 87% 82%

2 9 11 5 13 18 6 3 9

11% 22% 18% 26% 32% 30% 22% 13% 18%

19 21 40 15 21 36 15 31 46

70% 91% 80% 58% 91% 73% 79% 76% 77%

8 2 10 11 2 13 4 10 14

30% 9% 20% 42% 9% 27% 21% 24% 23%

41 60%

27

40%

42 62%

26

38%

39 57%

29

43%

Suggest Documents