Refining Grammaticalization Workshop

This is another reminder of a great opportunity to (further) acquaint yourselves with the wondrous world of grammaticalisation. And where else could it take place than in the loveliest of lovely cities in Europe? That is entirely correct: BERLIN!

http://www.geisteswissenschaften.fu-berlin.de/en/v/Refining_Grammaticalization/Programme/index.html

The programme is stunning with its variety of brilliant speakers. There is only one message to take home from this post: go and register.

Hope to meet lots of people there, maybe even you.

Denies

‘Possibly’ means that we’re not going to the game.

Much research has been done on modality and modal adverbs. As to the epistemic domain, it is often said that ‘possibly’ denotes a 50% likelihood that something is true and that ‘probably’ expresses high likelihood. As such, ‘possibly’ has been related to the modal verb ‘may/might’. While watching a film (The Pursuit of Happyness) the other day, I could not help but be fascinated by the following dialogue (bold is emphasis in speech, cursive are the modal markers):

Father: And maybe we’re going to the game.

Son: Where are we going now?

Father: Just to see someone about my job.

Son: I don’t understand.

Father: You don’t understand what?

Son: Are we going to the game?

Father: I said possibly we’re going to the game. You know what possibly means?

Son: Like probably?

Father: No, probably means there’s a good chance that we’re going to the game. And possibly means we might, we might not. What does probably mean?

Son: It means we have a good chance.

Father: And what does possibly mean?

Son: I know what possibly means.

Father: What does it mean?

Son: It means that we’re not going to the game.

This little dialogue is a beautiful reflection of the subtle nuances of the English modal system, where ‘maybe’, ‘possibly’ and ‘might’ are linked and opposed to the present progressive tense and the adverb ‘probably’. Even more, where ‘possibly’ is related to the negative. I don’t have that much to say about this – because I should actually be writing a paper – but thought it was a particularly interesting bit of the film to share with you interested readers!

‘Better start work.’ – ‘Yes, you better had.’ — an alternation study.

1. Introduction

The current post investigates motivations for the alternation pattern between the comparative modal constructions had better (1) and better (2) (van der Auwera and De Wit 2010, Denison and Cort 2010, Mitchell 2003, Colllins 2009).

(1) `If you know where she is, you had better tell me at once …’ (BB-cF022211)

(2)       ‘ … I won’t be working for you either. We’ll be working together. You better be straight about that from the beginning.’ (BB-cF022152)

The post is meant as a more robust and statistically supported continuation of the pilot study conducted by van der Auwera and De Wit (2010). These authors tentatively suggest that the researched deontic expressions of advisability or optativity can be distinguished from each other in a number of ways, viz. 3rd person Advisee is animate vs. inanimate, UK vs. US, type of subject (pro)noun as Advisee, conversational vs. not conversational.

2. Hypotheses

In summary, the following hypotheses can be highlighted.

Hypothesis 1:

had better is confined to narrative data, whereas better is to conversational data. The latter is especially so for UK English.

Hypothesis 2:

had better correlates with the Advisee (3SG/PL), whereas better correlates with the Advisee (1SG, 1PL or 2SG/PL).

Hypothesis 3:

better is not more characteristic of UK English than US English.

3. Methodology

3.1 Preliminaries

In the following, the alternation between had better and better will be further investigated. I have left ‘d better out of consideration for two main reasons. Firstly, simple and multiple logistic regressions of the kind used here are conducted on data with binary outcomes. Second, better and had better are furthest removed from each other in both frequency (at least in my dataset) and – according to van der Auwera and De Wit –  in use (with better being largely confined to third persons and conversational data). Another reason adding to my decision to leave out ‘d better is to keep this study as straightforward as possible.

I used the Collins Harper Online Wordbanks Corpus (553 171 489 tokens) to investigate written and spoken UK and US English. Again to keep the study straightforward, I have tried to keep my sample small but representative: Firstly, I selected only written material from magazines and from books. Spoken data contains various unspecified genres. Secondly, all data ranges from 2002 to 2005, a set which I believe will highlight the synchronic condition of Present Day English. I found 71 instances of better and 243 instances of had better.

This study is primarily quantitative. As mentioned above, I have conducted logistic regression tests in the statistic program R. Overall, a logistic regression test provides an ‘analysis and prediction of a dichotomous outcome’ (Peng, Lee and Ingersoll 2002, p. 3). In this paper, I used two popular commands to do this: a glm (generalised linear model) for the main part of the analysis and an lrm (logistic regression model) which offers more insight in the fit of the model and its predictive power. I have tried to conform to the conventions set out by Peng, Lee and Ingersoll (2002). My data consists of 71 deontic sentences with the short better + Infinitive construction and 243 deontic sentences of the longer had better + Infinitive construction (cf. (1) and (2)). The success level or outcome variable (Obsid) is had better, the failure level is better. In the light of testing van der Auwera and De Wit’s hypotheses (cf. supra), I chose the following predictors of which they tentatively suggested that they could be significant: Country (UK or US), Conversational (Yes or No), Advisee (1SG, 1PL, 2SG/PL or 3SG/PL) and Animacy (Yes or No)[1].

3.2 Methodological issue

A methodological issue was brought to my attention after conducting the glm command for a first time. The predictor Animacy is actually only relevant with respect to third person advisees, because, as mentioned, the other persons can only be animate (presuming we do not personify objects and refer to them as ‘you’, but I did not find that in the data anyway). I thus had to compute a combined predictor (Advisee2) for Animacy and third person Advisee using the commands in Appendix 1. When applying Advisee2 in another glm (fit2), I found that only the predictor Country and the combined predictor Advisee23SG/PL/Animate were significant (Appendix 2). Another model (d.glm), omitting Advisee/Advisee2 altogether, showed me yet other results: this time, all of the predictors were significant. This indicated that the predictor Advisee (and Advisee2 for that matter) was not decisive for the alternation pattern between better and had better. An ANOVA-test testing the different fits of the two models indicated that continuing with a model without the Advisee(2) predictor (d.glm) would be better. The drop1-measure (a Chi square test) further confirmed that all the predictors in my second model were significant (Appendix 3).

4. Results

4.1 General

1 van der Auwera and De Wit’s hunch about three of their predictors was correct. What the output shows about the predictors, however, is different from the suggested tendencies (treated in section 5.1). The output of d.glm (Appendix 4 and table 2 and 3 below) reveals that all of the predictors were significant with Country p<.05 , Conversational p<.05, Animacy p=0.

As already mentioned, the success level is had better. The estimates reveal the following findings:

1. There is a low probability for US English to use had better because the estimate of CountryUS is roughly -0.61.

2. There is an equally low probability for conversational instances using the had better variant because the estimate of ConversationalYes is -0.63.

3. There is a very low probability for the subject (pro)noun to be inanimate when using had better because the estimate is -1.17.

A first model diagnostic can at this stage be made: the residual deviance (310.96 on 310 degrees of freedom) shows that there is no sign of under- or overdispersion. The residual deviance does not differ much from the degrees of freedom and this confirms that the model has a good fit.

2 From the output of the lrm, we derive that the deviance of the model with 3 predictors is significantly smaller than the intercept only model because p » 0.  The likelihood ratio chi-square (L.R.) is not very high (24.73), which also means that the fitted model is a better fit than the intercept only model. C and Dxy assess the classificatory quality of the model (Speelman 2011). My model does not have ‘decent’ classificatory quality because Dxy (.424) is lower than .6 and C (.36) is lower than .8.

3 One of the many good things about logistic regression is that you can distinguish partial effects, which are the effects of an individual predictor when controlled for the other predictors (see Appendix 5). This gives a much clearer view of the actual effect of a predictor. For my data, the partial effect commands revealed that (1) the UK had a probability of roughly 85% that had better will be used when controlled for Animacy and Conversational, (2) the UK had a probability of roughly 85% that had better will occur in non-conversational instances and that (3) the UK has a probability of roughly 75% that the subject (pro)noun will be animate when using had better.

5. Discussion and conclusion

5.1 (Dis)confirming hypotheses

Hypothesis 1: had better is confined to narrative data, whereas better is to conversational data. The latter is especially so for UK English.

When reiterating the above finding 2 that conversational instances are not likely to occur with had better, we see that the first hypothesis is confirmed (Conversational p<.05).  We can also see that, proportionally, UK makes more use of better than US in conversational data. An explanation for this confinement to conversational data could be the level of informativity that is associated with better constructions.

Hypothesis 2: had better correlates with the Advisee (3SG/PL), whereas better correlates with the Advisee (1SG, 1PL or 2SG/PL).

Section 3.2 explained that the Advisee predictor was not significant for the prediction of one of the studied variants. When addressing the hypothesis then, it becomes apparent that compared to US English UK English prefers the use of 1SG and 1PL with had better and that the use of 3SG/PL/A+IA is preferred in the US with better, whereas in UK English there is a more balanced distribution. The partial effects test (section 4.1) also confirms that the UK has a probability of roughly 75% that the subject (pro)noun will be animate when using had better.

What d.glm did reveal, however, was finding 3. That is, there is a very low probability for the subject (pro)noun to be inanimate when using had better. This might seem counterintuitive seeing the higher frequency of had better. However, it could be indicative of a specialised use of better with inanimate subjects. Further research and a larger set of data will have to be conducted to get to the core of this.

Hypothesis 3: better is less characteristic of UK English than US English.

The predictor Country was a significant predictor and indeed it predicts that better is more characteristic of US English. This was already shown by finding 1 above . In the mosaic plot, which is sadly not available to you, the area covered by had better for UK is larger than for better in comparison to that covered for US. The reason for this might lie in the assumption that the better constructions have been used for a longer period of time in US English than in UK English. A more widespread distribution would then not be surprising. This claim, yet, is not mine to make because I do not have access to data that would confirm it.

5.2 Conclusion

I conclusion, van der Auwera and De Wit’s pilot study had the right intuitions about the modal comparatives had better and better. My larger dataset showed that there is indeed a difference among UK and US English with a preference for better for US English. Furthermore, conversational data was witnessed to correlate with the use of better. In one respect, however, van der Auwera and De Wit’s analysis turned out to be irrelevant for the choice between the two variants. My analysis illustrated that the grammatical number of the subject of the sentence is in fact not significant for the choice between either variant. Instead, it is the predictor Animacy which is most determining: there is a very low probability for the subject (pro)noun to be inanimate when using had better.

References

Biber, D., Johansson, S., Leech, G., Conrad, S. and Finegan, E. 1999. Longman grammar of spoken and written English. London: Longman.

Collins, P. 2009. Modals and quasi-modals. Amsterdam & New York: Ropodi.

Denison, D. and Cort, A. 2010. Better as a Verb. In: Davidse, K., Vandelanotte, L. and Cuyckens, H. (eds.). Subjectification, Intersubjectification and Grammaticali-sation. Berlin, New York: Walter de Gruyter, pp. 349-283.

Hopper, P.J. and Traugott, E.C. 2003. Grammaticalization. Cambridge: Cambridge University Press.

Leech, G. 2003. Modality on the move: The English modal auxiliaries 1961-1992. In: Facchinetti, R., Krug, M. and Palmer, F. Modality in contemporary English. Berlin & New York: Walter de Gruyter, pp. 225-240.

Mitchell, K. 2003. Had better and might as well: On the margins of modality. In: Facchinetti, R., Krug, M. and Palmer, F. Modality in contemporary English. Berlin & New York: Walter de Gruyter, pp. 131-149.

Peng, C.J., Lee, K.L. and Ingersoll, G.M. 2002. An Introduction to Logistic Regression Analysis and Reporting. The Journal of Educational Research 96(1), pp. 3-14.

Speelman, D. 2011. Logistic Regression in Corpus Linguistics.

van der Auwera, J. and De Wit, A. 2010. The English Comparative Modals. In: Cappelle, B. and Wada, N. Distinctions in English Grammar. Offered to Renaat Declerck. Tokyo: Kaitakusha, pp. 127-147.


[1] This is for obvious reasons only relevant for third person subjects. Inanimate (IA) will only refer to 3SG/PL.

Refining Grammaticalisation: Conference in Berlin

I am already looking forward to the following conference in Berlin (February 2012). One of my teaching professors brought this to my attention. Big names such as Brian D. Joseph, Hubert Cuyckens and Ulrich Detges will be speaking there, so I would urge everyone who wants to stay up to date in the ‘grammaticalisation business’, who has the money to go (government where are you?), the time to invest in it  (next to dinner parties and inbetween essay-writing), and who is not entirely tied to their computers to go!

Check out this blog

This is a blog about ‘language, its evolution and anything else in-between’. Do have a look at it, because it is genius.

http://www.replicatedtypo.com/

Modal Particles in English? Certainly Maybe! The Abstract

In recent years, an increasing emphasis on dynamic models of language change has become popular amongst scholars. Grammaticalisation theory offers a framework for orienting grammatical language change, whereby lexical meanings become grammatical or less grammatical meanings become more grammatical. In an attempt to disconfirm Hoye’s (1997) claim that the modal adverb certainly is a grammaticalised, reinforcing modal particle in collocation with modal verbs, this synchronic corpus study draws on the parameters proposed by Lehmann (1982[2002]) and criteria mentioned by Bybee (2003) and Hopper and Traugott (2003). The findings point at a slight inclination towards grammaticalisation of the modal adverb in mid-position, i.e. between the modal auxiliary and the main, lexical verb. Parameters such as cohesion, scope, syntagmatic variability and paradigmatic weight display these inclinations, supported by a quantitative analysis (frequency). Other parameters, viz. paradigmaticity and paradigmatic variability, have shown that certainly is far from grammaticalising. Indeed, a synchronic study can only reflect diverging findings of items undergoing language change (‘layering’) and thus this study can only be suggestive. Overall, although certainly might show signatures of grammaticalisation, it is maintained that it should not be viewed as a modal particle.

Modal Particles in English? The results…

In this second part on my MA dissertation, I have decided to render the results with limited discussion, partly because last post was just too long for blogging conventions, I feel; and partly because the results are interesting but not ground-breaking nor counter-intuitive. I could be very brief and show my abstract again, but maybe a little bit more explanation is required.

Conclusion

The days of absolute truth-finding are over, and I would not want to proclaim anything like truth. What I have found is suggestive of trends, but has been subject to errors in the online corpus – my study has had its limitations. In a world of superstition and paranoia, this presents the subtle work of irony to show that indeed we can never fully know what is going on (in language). But here, then, is how I make this study meaningful.

The data has revealed that certainly in mid-position as a reinforcing adverb with subjective, abstract meaning appears with all modals with a high frequency. As such, it is a plausible candidate for grammaticalisation (Bybee 2003; Hopper and Traugott 2003). Although certainly tends to appear most often with would and will, I believe that, really, the broader picture is most important for determining its status. The fact that it combines with all modals in mid-position fulfilling the same function is indicative of its spread and this ability is vital for grammaticalising constructions; only when there is high frequency in a range of syntactically similar environments, can a construction be propagated and eventually fixated (Croft 2000). Indeed, following Noël (2007), Hopper and Traugott (2003), Bybee (2003), I believe that it is the construction in which certainly appears that has lent to it its status as a potential modal particle as Hoye suggested (1997, p. 210). As I have tried to show – elaborating on Hoye’s claim – the adverb seems very integrated into the verbal phrase. Certainly is more syntactically ingrained (exemplified by the high frequency) in the ‘middle field’ (Traugott 2007), and it co-occurs with all modal verbs in the same reinforcing, subjective function. There is a distribution of labour in that it links to the modal auxiliary and adds to the propositional content. This aspect of certainly should not be confused with the other uses it can have. ‘Layering’ is a key trait of grammaticalisation and allows for divergence. The lexical meaning has been relegated to pre-mid-position.

The six parameters have suggestively pointed at grammaticalisation up to a certain degree, but I am not entirely convinced. In favour of the claim of grammaticalisation are the syntagmatic cohesion, the paradigmatic weight and the scope. The syntagmatic cohesion has proved highest in mid-position (frequency), the scope has increased (certainly occurs with all modals) and the paradigmatic weight is low in mid-position (subjective, reinforcing meaning): There is a meaning difference between certainly in mid-position and certainly in pre-mid-position which can be allocated to its syntagmatic position. In mid-position, an abstract, reinforcing meaning can be discerned next to subjectivity, whereas in pre-mid-position the meaning has retained much of its concessive-contrastive use and whereby it approximates the use as a sentence adverb. The functional syntagmatic variability is low but not non-existent; certainly is still open for modification and this is an argument against it being a modal particle. The paradigmatic variability is high overall and the paradigmaticity is low.

To link these findings to the properties of modal particles, I have made a checklist which shows the following:

checklist

Overall, I maintain that my hypothesis is confirmed seeing that only a few parameters have shown an inclination towards grammaticalisation (cohesion, functional syntagmatic variability, paradigmatic weight, scope; frequency). Even with these signatures of grammaticalisation, I do not consider certainly as a modal particle for the reason that it allows for modification (e.g. will almost/most certainly). Probably the most deciding factors for my non-inclination to accept certainly as a modal particle are

›          English’s lack of modal particles overall

›          a persisting association of certainly with its adverbial use.

If English has for years been understood as a language without particles and we find items like certainly which do not entirely match the traits, then I think it is safe to assume that English still does not have what is commonly understood to be a ‘modal particle’. I could postulate that in the unmodified cases, certainly has a different, abstract meaning, which would support a theory of grammaticalisation of certainly in mid-position. The fact that a cumulative reading is not available for certainly in mid-position, additionally, would also be in support of grammaticalisation.

Although I have emphasised that intuition should not be used to make claims, my intuition has led me to question Hoye’s claim and it is my intuition – after this study – which remains sceptical of his suggestion.

An important remark is meant to relativise the above findings and claims. Also outside of the realm of modal synergy, certainly also occurs in mid-position most frequently. This could mean that it functions as a reinforcing adverb in mid-position between any auxiliary and main verb and that this function is not linked to modal synergy. Where it originated (in modal synergy or not) is a question which this study cannot answer and diachronic research will have to point out, but the changing semantic and syntactic nature of certainly is an interesting topic which needs further exploration in order to get a more complete understanding.

Modal particles in English? Certainly Maybe

This reduction of my MA will be published in two parts. Part one introduces the reader to the subject matter, pointing out the research question and hypothesis and the methodology. Part two discusses the results and findings to then conclude critically. The bibliography will become available in the second post.

Part one is accompanied by some great music produced by my younger brother, Kristof.

Intense Meditative Trip Sessions Vol.1 by The Mixing Chef

1. The seeds
My first MA dissertation was a corpus study, concerned with finding out the grammatical status of certainly in collocation with modal verbs. The seeds for the dissertation were sown by Hoye in his book Adverbs and Modality in English (1997). His book investigates the possible collocations and their meanings and uses of modal verbs and modal auxiliaries, also called modal synergy by Hoye – a research area which has only recently regained its fervour (e.g. Geurts 2006, Huitink and Geurts 2008, Zeijlstra 2007, Anand- Brasoveanu 2010) since Halliday (1970), Lyons (1977) and Hoye himself (1997). The following is an example of such a combination:

(1) Could I possibly borrow your history book?

Hoye’s book further offers pragmatic reasons for using modal synergy and orientates it in Speech Act Theory. A few of his claims (pp. 208-212), seemingly impulsive, attracted my attention and left me in deep pondering. Page 208 introduces a question which I wanted to answer in my dissertation: ‘Modal adverbs or modal particles?’. Hoye suggests that certain modal adverbs have become ‘grammaticalised’, ‘delexicalised’, ‘lexically redundant’ in the environment of modal verbs and argues for the name ‘modal particle’ for a subset of what he calls reinforcing modal adverbs. He only very briefly sums up a few reasons (viz. syntax and semantics) but he does not engage in offering definitions and further explanations as to the status of these modal adverbs as modal particles. Certainly is one of those adverbs, and in my dissertation it was my aim to investigate whether Hoye’s intuitions about it were true. Due to word constraints, I was only able to look at certainly in collocation with will, would, may, might and must. Before I go into the details of my findings, I will first say a few words about certainly.

Certainly in Modal Synergy
Certainly is a modal adverb related to the adjective certain and the noun certainty. In adverbial use, it thus expresses high certainty, but it also has concessive-contrastive meanings. Certainly has strong modal strength (it expresses high likelihood) and as such it always reinforces what it modifies. In sentence-initial position, certainly is a high probability adverb which expresses ‘the speaker’s strength of conviction in the truth of the adjoining proposition’ (Hoye 1997, p. 190); it reinforces the speaker’s belief. The default cases of modal synergy are called modal harmony, whereby the modal adverb and the modal verb are of the same modal strength. When certainly functions concessively (very often followed by a but-clause) (Swan 2005), the collocational restrictions of modal harmony do not always hold and certainly will appear with modal verbs which express a lower degree of likelihood. Cp.

(2) Certainly he must attend classes.
(3) Certainly he could be aggressive, but I have never seem him act it.

That being said, certainly actually readily collocates with all modal auxiliaries. Also Simon-Vandenbergen and Aijmer 2007 (pp.89-90) confirm this and add that it most frequently co-occurs with would and will. These uses, they further say, are epistemic (predictive) or dynamic (volitional). When combined with epistemic modals of possibility, certainly is said to express concession; but other than that it would convey ‘straightforward affirmation of truth’ (Hoye 1997, p. 162).

Now I will briefly introduce how in a theory of grammaticalisation, I went about conducting the study.

2. The core concepts
Grammaticalisation
Grammaticalisation is to be situated in theories of language change. It is a type of change which in itself comprises several processes of change. As the name lets on, the change with its inherent processes has to do with language items become part of the grammar. However, next to a syntactic change, grammaticalisation is also a ‘global change affecting … the morphology, phonology and semantics’ (McMahon 1994, p. 160).
Also Hoye has attempted at pinpointing what seems to him a grammaticalised adverb: viz. certainly. In what follows, I provided an introduction into grammaticalisation. The aim was to offer a comprehensive overview of grammaticalisation in order to (dis)confirm later on Hoye’s intuitions about certainly, which he left remarkably undefined or backed up with proof from the used corpus.

Definition
The traditionally quoted definition from Hopper and Traugott (2003, p. 18; my emphasis) goes: Grammaticalisation is the ‘change whereby lexical items and constructions come in certain linguistic contexts to serve grammatical functions and, once grammaticalised, continue to develop new grammatical functions’. This change is to be understood as a gradual continuum, a slippery slope of unidirectionality, which can consist of several terminals, i.e. a construction need not follow the cline to the ‘end’, but can halt its development at any point on the continuum. Triggers for grammaticalisation are multifaceted but far beyond the scope of this study; to name a few: sociolinguistic pressures, principle of economy, expressivity enhancement etc. (Croft 2000, de Haan 2005, Hopper and Traugott 2003).
For the purposes of the current study, I would like to highlight the following quote from Hopper and Traugott (2003, p. 2; my emphasis, square brackets from p. 18): ‘Grammaticalization … is part of the wider linguistic phenomenon of structuration, through which combinations of forms may in time come to be fixed in certain [grammatical] functions [in certain linguistic contexts]’. This process is said to be marked by phonological reduction, semantic bleaching, (a/) pragmatic shift(s), decategorisation , and syntagmatic and paradigmatic restrictions. For example, the verb have has over time come to be used as a temporal (‘possessive perfect’) and a causative auxiliary next to its original lexical main verb use of ‘possession’. The above quote goes back to the idea that one cannot study the change of items by themselves but that the context is important for the semantics and the syntagmatic properties of that item (cf. Firth; Bybee 2003; Bybee and Cacoullos 2009; Noël 2007). In an attempt to marry grammaticalisation theory to construction grammar, Dirk Noël (2007) explains that the context of a linguistic item is a vital prerequisite for eligibility for grammaticalisation (see also Diewald 2006) and that the elements in this context should be highly productive, i.e. have a high frequency (cf. Bybee 2003).
An important feature of grammaticalisation or grammaticalised constructions is LAYERING (Hopper and Traugott 2003, pp. 124-126). This means that several older and newer lexical and grammatical meanings of one form can co-exist: The grammatically reanalysed construction has come to serve functions independent of the previous – less grammatical – construction. This, however, does not necessarily mean that the former analysis of co-occurring elements is lost or that the items of the co-occurrence cannot exist in other constructions with their original – lexical – meanings. Indeed, an older meaning can be said to persist. This co-existence of several meanings is called layering – a phenomenon of context-dependent polysemy. Layering can allow for what is called divergence, where some meanings diverge into a certain direction with higher frequency and the others lagging behind. I recall an example of layering from the opening paragraph: to have.

(4) I have food. possessive, lexical ‘have’
(5) I have eaten. possessive, perfective ‘have’
(6) I had him do it. causative ‘have’

In summary :
WORKING DEFINITION FROM TORRES CACOULLOS AND WALKER (forthc., p. 1):
Grammaticalization is the set of gradual processes, both semantic and structural, by which constructions [(collocations) next to single items] involving particular lexical items are used with increasing frequency and become new grammatical constructions, following cross-linguistic evolutionary paths.

Lastly, to link in with another important aspect of language change: in the diachronic take characteristic of her work, Traugott (1995b) suggests that, integral to grammaticalisation, the items in the process of language change undergo a semantic change along the following cline:

objective > subjective > intersubjective

This process can be called (inter)subjectification, ‘a gradient phenomenon, whereby forms and constructions at first express primarily concrete, lexical, and objective meanings come through repeated use in local syntactic contexts to serve increasingly abstract, pragmatic, interpersonal, and speaker-based functions’ (Traugott 1995b, p. 32; emphasis mine). The modal verbs, for instance, have been noted as instances which have become grammaticalised from lexical to deontic to epistemic auxiliaries on the one hand and from objective to subjective on the other. Especially with respect to deontic modals, a cline from ‘objective’ universal obligation to subjective, individual obligation is observed (Traugott and Dasher 2005, p. 114). Some modal verbs are also intersubjective. Indeed, Palmer (1986, p. 16) puts it ever so poignantly: ‘Modality could … be defined as the grammaticalization of speaker’s (subjective) attitudes and opinions’. This is an interesting choice of words; especially seeing that Traugott discerns a semantic process ((inter)subjectification) within the grammatical one (grammaticalisation).

What does it mean for an item/construction to be ‘grammaticalised’?
As with any theory, the boundaries of what is defined need to be delineated. For grammaticalisation, Christian Lehmann designed six parameters to map the grammaticalised status of a linguistic item. The following is based on Szczepaniak (2009), Diewald (1997) and Lehmann (1982[2002]). Lehmann considers the autonomy of a linguistic item as the key criterion to distinguish grammaticalised or grammaticalising constructions from lexical items. The more autonomous a form is, the less grammaticalised it is. He divides this concept of autonomy up into three aspects which indicate the degree of grammaticalisation: Weight, Cohesion and Variability. In order for an item to be autonomous, Lehmann argues, it has to have considerable semantic and formal weight. With respect to the second aspect, cohesion, Lehmann proposes that the autonomy of an item decreases up to the extent where it systematically combines with other items. Cohesion is thus the systematic combination of items and increases in grammaticalisation. The third aspect, variability, is an indicator for how an item can be moved around whilst being in a syntactic relation with other items. High variability is characteristic of open-class items which can be easily replaced by other semantically related items and correlates with high autonomy; e.g. nouns are highly variable. Highly variable items consequently have more semantic and phonological weight and they are less -or not at all- grammaticalised. The other end of the spectrum is obligatory use of an item; e.g. will or going to to express future tense. Lehmann divides these three aspects up once more according to a syntagmatic and a paradigmatic axis, which represent horizontal and vertical choices respectively available to the language user. Visualised, these axes look like this (Szczepaniak 2007):

paradigmatic and syntagmatic axes sczcepaniak

Figure 7: Paradigmatic and Syntagmatic Axes

The below table lists these six parameters along the three criteria – weight, cohesion, variability – and the two axes – syntagmatic and paradigmatic. The arrows show the respective in- or decrease of the traits in grammaticalisation.

6parameters

Table 2: Parameters for Grammaticalisation: Lehmann 1982 [2002]

I will not go into further detail about the traits of grammaticalisation.

3. Research Question and Hypothesis

Research Question: Is certainly in collocation with a modal verb (will, would, may, might, must) a grammaticalised modal particle from a synchronic point of view?

My hypothesis is that the adverb is not fully grammaticalised even though it might show signatures of grammaticalisation. As could be expected, modal particles can be the result of grammaticalisation and as such, they exhibit the properties inherent to grammaticalisation in terms of syntax, phonology and semantics.
Following Noël (2007), Bybee (2003) and Hopper and Traugott (2003) grammaticalisation affects a construction primarily and a single word secondarily; I suggest that, for modal synergy, a structural unit is formed of a modal verb and an adjacent modal adverb in mid-position, e.g. would certainly, must certainly etc. Mid-position is the ‘natural habitat’ of the modal particle and if there is grammaticalisation of certainly into a modal particle, this is consequently where we would expect to find it. Moreover, if this were a grammatical unit/construction consisting of two grammatical constituents, the grammaticality would lie in the bondedness (syntagmatic restriction) of the two elements, and the semantic and paradigmatic restrictions which are said to be part of grammaticalisation (cf. Lehmann’s parameters): we would expect an abstract meaning and perhaps reduced phonological properties (which I cannot test), paradigmaticity, low paradigmatic variability and high cohesion with modal verbs in general. Scope is a contested parameter and it seems that in this case too, we will deal with increased scope. Lastly, as Bybee (2003) indicated, frequency plays a staple role in the propagation of an item to becoming grammaticalised (see also Croft 2000).
Additionally, if there is such a thing as subjectification, and we assume that Hoye’s claim about the grammaticalised status of certainly is correct, then certainly has to have a more abstract, subjective meaning. This has to be the case, seeing that processes of grammaticalisation entail subjectification (Traugott and Dasher 2005).

4. Methodology

I used the Collins Harper Online Wordbanks Corpus (553 171 489 tokens) to investigate written and spoken UK English instances of modal synergy with certainly. The texts from which I found these instances range from 1990 until 2005, a representative set of contemporary trends in present-day English. This study thus adopts techniques from corpus linguistics, but moreover draws on a theory of language change: grammaticalisation.
Hoye (1997, p. 185) refers to possibly, probably and certainly as the central modal adverbs. A study dedicated to one of them, viz. certainly, seems justified from the point of view that being a central modal adverb, a higher frequency and thus statistically more feasible result will be found. The main aim of this study is to comprehensively render an account of the (grammatical) status of the modal adverb certainly in collocation with modal verbs according to Lehmann’s parameters discussed in the literature review. Issues of lexical weight, bondedness, paradigmatic variability, structural scope and functional syntagmatic variability of the items in modal synergy will thus be discussed in succession. The paradigmatic, phonological weight will be outside the realm of the study because I have not had access to recorded data and so stress and intonational patterns were not noted. Paradigmatic, semantic weight will be tested along positions in the sentence. For example, the difference between these sentences will be discussed:

(7) He certainly will come.
(8) He will certainly come.

Bondedness is largely determined by frequency and phonological reduction, the latter, again, was not available for analysis to me. I determined the paradigmatic variability by comparing the frequencies of the modal adverbs which collocate with all modal verbs. Highly frequent occurrences are more likely to be integrated into a paradigm, low frequency reflects bigger autonomy and more choice for the language user.
The discussion about the semantics of the adverb will be linked to (inter)subjectivity. More statistically then, frequency is a first indicator of, and simultaneously an enabling factor for, grammaticalisation. Increased frequency of a construction can make it open for grammaticalisation and a grammaticalised construction appears more frequently. I call this ‘mutual reinforcement’. Frequency distributions will be useful to offer an insight in synchronic trends and they will introduce us to raw patterns in the data.
Due to time and word constraints, I was only able to discuss certainly in collocation with five modal verbs: will, would, may, might, must. Certainly is said to combine with will and would most frequently (Simon-Vandenbergen and Aijmer 2007, pp.89-90) and thus an analysis of these varied instances will offer a better understanding of the general behaviour of the adverb. The other modals, however, operate on the other end of the spectrum and infrequently – so do not combine not at all – combine with certainly. Mirroring the highly frequent combinations to the infrequent ones will shed light on whether there are any meaning differences. If there are, then certainly serves different functions in different collocational combinations; if there are none, I can conclude that the meaning of the adverb is largely consistent and this could give me reason to postulate that certainly is persisting in other categories as well.
The combination of certainly and these modal verbs were studied in two syntagmatic positions only; the reason for which is that I noticed a considerable difference in meaning between them which led me to think perhaps there was something special going on. I distinguished mid-position of the adverb, e.g. He would certainly come, and pre-mid-position of the adverb, e.g. He certainly would come. Additionally, I am only analysing adjacent constructions, so phrases like ‘it would very certainly be the case’ are disregarded. The reason for that is that adjacent constructions are more easily subjected to a successful analysis, are more frequent and thus initially more open for grammaticalisation (the innovative stage).
(Complete) grammaticalisation of a construction implies that the construction involved as a whole expresses a specific grammatical meaning. This goes hand in hand with a certain semantic-pragmatic meaning or implication. If in my data I find a high frequency of particular syntagmatic structures with one and only one grammatical meaning (and a corresponding pragmatic-semantic implication), this would be indicative of a grammatical(ising) construction. It would be even more so, if the same elements involved in a different syntagmatic or paradigmatic constellation render another meaning. In order to find out whether the sentence meaning changes according to the constellation of the modal synergy, I will discuss syntagmatic and paradigmatic patterns of co-occurrence and their meanings (cf. Lehmann’s parameters). For this synchronic study, it is important to keep in mind ‘layering’ (i.e. one word/construction in the process of language change can have several meanings simultaneously). Thus, a quantitative (frequency) embracing a qualitative (semantic-pragmatic) analysis will offer the clearest representation of ongoing tendencies.

Limitations

A few important remarks with respect to my study need to be mentioned. Firstly, it is extremely difficult, if not impossible, to exactly determine where the scope of a (modal) adverb lies (Hoye 1997, Sweet 1892). And even if the scope is claimed to be identified, hard and fast proof for these identifications is mostly absent or flawed. However, what seems to be agreed on is that modal adverbs either modify an entire clause (sentence adverbs), in which case they are lexically more autonomous, or a part of the clause. In any case, and as Sweet (cf. supra) has indicated, the verbal phrase can always be subject to modification by modal adverbs. The modal verb is thus always within the scope of the modal adverb.

Furthermore, limited within the confines of an online database which was –unfortunately – occasionally subject to errors and crashes, my observations are suggestive of trends, but by no means are they exclusive or exhaustive. One of those limitations was the inability of the corpus to display contracted forms at times, so that I never really ended up with the complete set of contracted forms such as ‘d or ‘ll certainly. This left me unable to distinguish between contracted had and would on a micro-level, but on a macro-level it disabled me to finely tune my analysis to more subtle differences of language use. With contracted form followed by certainly, we would for example expect that – because of the phonologically reduced nature of the auxiliary – the stress patterns are more geared towards certainly, thereby putting more emphasis on the adverb. Further research with more comprehensive data and tools is needed to confirm and expand my findings.

Next to constraints resulting in a limited data-set and analysis, there are further limitations to this study. I will not be able to take into account modal prosodic features such as the fall-rise (Perkins 1980). Although these are not grammatical in any way – English is not a tone language – there are patterns of prosody which most frequently correlate with epistemic modality. Other than that, other pragmatic or emphatic stress patterns could not be identified, hampering a better understanding of the sentence meaning and the context.

Lastly, because of time and word constraints, this is a synchronic investigation yearning for diachronic succession. However, as Traugott (2006) very rightly pointed out, any diachronic study can only be successful in succession of a detailed synchronic investigation. This justifies my synchronic study at least in the respect that it facilitates future research.