At present, I claim familiarity with the `Principles & Parameters Approach' (PPA) [also known as `Government & Binding' (GB), `the Revised Extended Standard Theory' (REST), or `the Minimalist Program'], Cognitive Grammar, Generalized Phrase-Structure Grammar (GPSG), Head-Driven Phrase-Structure Grammar (HPSG), Lexical-Functional Grammar (LFG), Relational Grammar (RG) [also known as `Arc Pair Grammar' (APG)], Role & Reference Grammar (RRG), and Word Grammar. As can be seen, my acquaintance includes relatively `formal' frameworks (e.g., GPSG, RG/APG, and at least to some extent PPA), more `functional' frameworks (Cognitive Grammar, RRG, and Word Grammar), and some that partake of both characters.
Different theoretical frameworks have different foci of attention. This is one of the main reasons i hope and expect that they should be able to complement each other, different frameworks focussing on different details and adding up to a composite picture that, ideally, proves vastly more illuminating than each framework's facet by itself. But in addition to different foci, different frameworks have different axioms and, resulting from that as well as from the different foci, different methodological approaches to a given problem.
One of my chief concerns is the extent to which training and experience in one particular framework can blind a researcher to the methodological possibilities of others. This is one of my main reasons for wanting to train my students in a variety of frameworks, the better to promote methodological flexibility. I have found arguments in the literature to the effect that framework A does a `better' job of accounting for a certain corpus of data than framework B, and frequently such arguments real boil down to tautological demonstrations that B doesn't share A's axioms, foci, or consequent methodological approach, without recognizing that B's different axioms, foci, and approach may lead to a very different account of the data that may, nevertheless, be just as elegant or otherwise satisfactory as A's, if not more so. This problem has been a major focus of my thinking since i was an MA candidate, and i hope very much not only to be able to pursue fruitful research in this area, enabling differing frameworks to communicate with and cross-pollinate each other better, but to teach an ongoing seminar on the subject.
As just one example among many, what does the meaning of `understand' have to do with that of either `under' or `stand'? A comprehensive etymological dictionary like the OED may be able to answer this question, but it's not something the average English user carries around as productive information. Furthermore, how come `put' means `cause X to be in location Y', `put up' (at least in one usage) means `provide lodging for X', and `put up with' means `tolerate X'?
For convenience, i will refer to these morphemes as `verbal particles'. I am very interested in investigating their grammatical and semantic behaviour diachronically and cross-linguistically, with a view to the following questions:
In most cases, what options exist are typically ranked according to preference, within the general speech community, within the usage of a particular individual, or within a particular register of language use. Most of us accept the first category without thinking about it much. As far as the second is concerned, we may simply accept that other people within our speech community don't have to talk exactly like us, and may have different grammatical preferences. As for the third, we tend to concentrate on those registers which we need to use in our personal and career lives, make sure we understand what usages are appropriate to each of these, and ignore the others as much as possible.
Literary authors writing dramatic or narrative fiction, however, by the nature of their work are often called upon to represent a wider variety of grammatical options than a typical individual language user would normally use. In many cases a writer, such as Dickens, will use a variety of dialectal idiosyncracies in order to define the regional or class background of his or her characters; such exploitation is particularly characteristic of mystery fiction, in which such issues as regional, class, or occupational background can be of great significance. Or a writer, such as Shakespeare, may exploit grammatical options within a single dialect for the purpose of characterization by defining a range of variability and assigning to each of his characters a point along that range, so that each character's usage is completely consistent internally (at least with regard to the appropriate variables) but differs subtlely from that of all the other characters.
Where poetry exists, and it seems to exist in all human cultures, it co-exists with prose within a single speech community; we are no more likely to find a community whose members communicate consistently in verse than we are to find one in which they communicate solely in prose. Therefore, if current fundamental assumptions about linguistic science -- e.g., the specific innateness of the human language faculty, the equivalence of all `natural' languages, etc. -- are valid, then it stands to reason that within a given speech community the poetic expression native to that community must be generated by, and describable by, the same grammar that generates it prose expressions.
This is my basic working assumption. I hope to be able to pursue some research to demonstrate its validity. But there is an added dimension here that i believe makes the issue of great importance to syntactic research. While poetry and prose within the same community may be generated by the same grammar, i suspect that poetry is more likely to exploit the (exceptional) options allowed by that grammar. If this is true, then if one wants to know what the limits really are on what is grammatical in a given language it is necessary to examine verse data.
All languages use at least some of these, many languages use more than one, and i suspect that all languages in fact use all of them to some degree. But languages differ with regard to the relative importance each parameter has in the ultimate selection of one constituent-order option from a variety of possibilities. In English and many other Western European languages, grammatical relation is often the most important consideration; languages for which this is not the case tend to strike native English speakers, at least at first glance, as rather wild and unconstrained, and are often described in the literature as `free word-order'. But because the different parameters must be balanced and, if necessary, compromised, a certain amount of constituent-order freedom is manifest in all natural languages.
As i explained in a brief note in Natural Language and Linguistic Theory in 1991, languages vary not only in degree but also in kind of constituent-order freedom they exhibit. There are languages like Japanese that are commonly described as `free word-order' which are really `free phrase-order', in that they have few grammatical (or even perhaps semantic) constraints on the ordering of whole NPs or other phrases within the clause, but which seriously constrain the possibility of discontinuous phrases. And there are languages like Warlpiri or the early Indo-European languages that are truly `free word-order' in that they not only allow phrases to occur (almost) anywhere within the clause, regardless of their grammatical or semantic role, but they allow individual words that `belong' to a given phrase, in terms of semantic collocation and/or morphological agreement, to be `scattered' away from each other about the clause. One of my principal interests is the investigation of the various possibilities for constituent-order freedom exhibited among the languages of the world, with a view to questions such as the following:
I have since examined the issue repeatedly, from a number of different approaches, read a lot of literature on what i have come to call the `Dravido-Uralic Hypothesis', and consulted with many living (at least at the time i talked to them -- Robert Austerlitz has since died) specialists in relevant areas, both encouraging and sceptical. It is evident that some of the original reasons for positing a relationship between the Dravidian and Uralic families (e.g., the existence of vowel-harmony processes in Finnish, Hungarian, and Telugu) are inadequate or specious; but i have so far concluded that there are too many tantilizing hints of similarities for me to leave the hypothesis alone but that it will be damnably difficult to prove. Thus, it represents a major challenge, and i hope to spend a certain portion of my career time and energies investigating it thoroughly and possibly making a major contribution to the study thereof (and possibly also to the anthropological question of how two linguistic stocks so far removed in space could nevertheless be related!).
I should note that my interest in such a deep, long-range glossogenetic affinity has in some quarters earned me the reputation of being a Nostraticist, or at least being in the Nostraticist camp. I think i can (and have to) honestly say that this is not the case. I am obviously sympathetic to attempts to identify long-range glossogenetic affinity; i frankly don't see how any true historical linguist can fail to be. But my own personal fascination is merely with the hypothesis that Dravidian and Uralic are related, not that all the languages of the whole world, or even of the `Old World' are. And, as already indicated, my current belief is that it will prove a monumental task of comparative/reconstructive linguistic work to adequately demonstrate the truth, if any, of the Dravido-Uralic Hypothesis. Demonstrating the interrelatedness of all, or even most, of the recognized Old-World linguistic stocks would presumably require even more massive labour. And since the material one is dealing with -- the morpheme stock of any given language, or of a finite number of languages, and the length of individual morphemes in terms of phonological segments -- is essentially finite, it is clear to me that one is approaching a point of diminishing returns. The evidence upon which a case for the Dravido-Uralic Hypothesis might be based consists for the most part of an admittedly large number of rather small morphemes. Given the size of the morphemes (often only two or three segments apiece), one begins to have to be concerned about the possibility that even such a large number of apparent correspondences might be indistinguishable from chance similarity. This is one of the problems i'm going to have to wrestle with in the coming years if i'm going to pursue this topic. And, given this consideration, while i am sympathetic to the Nostracist goal i am not sanguine about its attainability. I am certainly not a partisan supporter of the Nostratic approach.
Languages in contact may compete with each other, a situation that may result in one language achieving hegemony over a community that had previously been either inherently multilingual or composed of several linguistically distinct monolingual communities. But it is not easy -- if, indeed, it is possible at all -- to predict which language will get ascendency; witness in the history of Great Britain the triumph of English over both the conquered Celtic-speaking population and the French-speaking Norman conquerors. Another possible outcome of a linguistically competitive situation is the definition of distinct spheres associated with distinct languages. In some cases this means merely using one language inside the home or within a narrow community while using a different one with people from outside, but in some such a strategy can develop in much more complex directions. In other cases of language contact, a new language may emerge as a sort of compromise between different languages, often combining elements from them. Such new developments are labelled variously `pidgins', `creoles', `lingua francas', or `koinés', dependeing on several criteria.
The dynamics involved in language contact situations typically involve not so much specifically linguistic factors as social ones such as prestige and ethnic pride. For me, it is in the interaction between such forces, and the details of how that interaction manifests itself linguistically, that the real interest in language-contact research lies.
At this time, i find myself particularly interested in investigating the history and current language-contact situations in the following geographical regions:
The following papers are also available from me. As they are much shorter than my dissertation, i am much more willing to make copies of them and circulating them at my own expense.
This short paper, originally written in 1986, discusses the ways in which the vocabulary of English has evolved to fit a variety of different countries around the world. In particular, it surveys some representative lexemes that in form are common to the English language as used in these countries but whose meanings have evolved in disparate ways in each of them, thereby raising the possibility of communication failure among people ostensibly using the same language.
This, basically a slightly revised version of a graduate-student term paper from 1987, brings together all the information i was able to find at the time about `passive constructions' in South Asian languages and about the challenges they present to formal syntactic theory. It includes some discussion of the methodological questions implicit in such a survey, in particular the problem of identifying what is, and what is not, a `passive construction'.
This paper, published in 1988 in Studies in the Linguistic Sciences, discusses the fact that the Vedic Sanskrit equivalent of what is commonly called `wh-fronting'
This 1990 paper discusses the syntax of the pre-verbal focus position in Vedic, arguing that a `downgrading' analysis whereby the element in focus is c-commanded by its own `trace', is the only analysis that meets the data.
This short essay, which was published at a `Topic...Comment' column in Natural Language and Linguistic Theory in August 1991, consists of a plea for the recognition by syntactic theorists of the distinction between `free phrase-order' and `free word-order' phenomena, and for the claim that they require independent syntactic processes to account for them.
This 1992 paper presents a complex argument, based on the ordering of constituents around the verb in verb-medial clauses in Vedic (normally a verb-final language), to the effect that Vedic grammar defined a constituent that included the verb and its complements but excluded the subject, and that the verb and its complements only rarely moved out of said constituent during the course of the derivation of a clause.
Examining a variety of clause-peripheral phenomena in Vedic Sanskrit, this paper, published in 1993 in Studies in the Linguistic Sciences, argues that they are not amenable to an analysis in terms of projections of a functional head Comp, but rather that such a head is best regarded as a syntactic coalescence of a variety of syntactic and pragmatic functions which UG allows some languages to keep syntactically distinct.
This 1993 paper argues that the ability in Vedic Sanskrit of single words to be fronted to clause-initial position even out of fairly deeply embedded phrases can only be accounted for if it is assumed that Binding Theory functioned somewhat differently in that language from the way it functions in English (i.e., there are inherent differences in the ways Binding Theory constraints can be satisfied in the two grammars), and in particular that the `Head-Movement Constraint' must have been null and void in Vedic, and therefore cannot be a linguistic universal.
This paper, published in the 1993 ESCOL (Eastern States Conference on Linguistics) proceedings, demonstrates that although the `Liberation' process presented by Geoff Pullum in 1982 to account for free word-order phenomena has some serious theoretical problems in GPSG, as demonstrated by Arnold Zwicky in a 1986 paper, these problems `magically' disappear when the process is reinterpreted within an LFG framework, with its different methodological assumptions and outlook, and thus that within such a framework Liberation is quite feasible as a means of accounting for free word-order phenomena.
This paper, presented at the 1993 OSU Workshop on Wackernagel's Law and published in the proceedings thereof, discusses the complex clause-initial string characteristic of Vedic Sanskrit and argues that its internal structure is the result not of syntactic parameters but of the interaction of independent phonological and semantic constraints. Includes the initial statement of the hypothesis that `discourse particles' are semantically vacuous and that their syntactic behaviour, at least in Vedic, may be partly due to this fact.
This paper, which has gone through several versions since 1993 and is currently being considered for publication in the Journal of English Linguistics, argues against the Pollockian hypothesis that syntactic and morphological processes are linked through the details of the structure of what used to be called `Infl' by showing that over the course of several generations in the middle of this millenium English either failed to behave in a manner consistent with this hypothesis or did so only optionally. It includes some discussion of the theoretical significance of such optionality and refutation of alternative views of the data (some of them formulated in response to earlier versions of this paper) that seek to salvage some version of the Pollockian hypothesis.
This paper, presented in 1994 at the Mid-America Linguistics Conference and published in the proceedings thereof, argues against the above-mentioned Pollockian hypothesis from a different point of view, showing that a variety of languages living and dead show evidence of defying the predictions of the hypothesis in various different ways, suggesting that the hypothesis is without real predictive value.
Note:Due to space constraints, the version of this paper that appears in the MALC proceedings is substantially shorter than the full version. Anybody who wants the complete version has to get it from me.
This paper, whose most recent version was presented at a poster session in Milwaukee in April 1996 and is currently in press but which dates back to a 1985 term paper, demonstrates that many arguments confidently presented in the RG literature for the `multistratal hypothesis' in fact fail to take into account the fact that a non-transformational framework like LFG has different ways of looking at the data and therefore can often account for it at least equally well without resorting to multistratalism.
And i continue to imagine courses i would like to teach someday, and to consider how such courses might best be organized and taught. I occasionally put in some time, not merely designing course syllabi and outlines, some of which can be accessed hence, but drafting actual lectures. After all, at least until i manage to land an academic job somewhere, i've got to live on my dreams.