Bay Area Typology Workshop

Program

8:30     Coffee, setup, test equipment, etc.

9:00     Welcome, introduction

9:15     Balthasar Bickel & Johanna Nichols (U.C. Berkeley)
           Wordhood revisited: Theory of types, theory of distribution, and theory of grammar

10:00   Elena Maslova (U. Bielefeld)
           On linguistic interpretation of quantitative cross-linguistic data

10:45   Coffee

11:00   Andrew Garrett (U.C. Berkeley)
           Synchronic vs. diachronic explanation of typological patterning

11:45   Larry M. Hyman (U.C. Berkeley)
           Compositional and/or templatic morphology: Suffix ordering in Bantu

12:30   Lunch

2:00    Judith Aissen (U.C. Santa Cruz)
           Harmonic alignment and morphosyntactic markedness in OT

2:45    Joan Bresnan (Stanford U.)
           Typology in Variation: An Stochastic Optimality Theoretic Approach

3:30    Coffee

3:45    Masayoshi Shibatani (Kobe U. and CASBS Stanford)
           The causative continuum

4:30    Dan Slobin (U.C. Berkeley)
           Typological differences in representation of motion events: Some cognitive and communicative consequences

5:15     Break

5:20 — 6:00 Round table

Abstracts

Judith Aissen
Harmonic alignment and morphosyntactic markedness in OT
An important set of implicational universals in morphosyntax depend on hierarchy alignment. These include the conditions for 'split ergativity', differential object marking, the choice of proximate in obviation systems, the factors in subject choice, and the realization of double object constructions. This paper reports on work which attempts to develop a formal OT account of these domains based on harmonic alignment, a formal operation for aligning prominence scales. I will sketch the basic technique, and and show how it can be successfully applied to some of the domains mentioned above.

Balthasar Bickel & Johanna Nichols
Wordhood revisited: Theory of types, theory of distribution, and theory of grammar
The status of formal theories (or theories of grammar), distributional (or quantitative) theories, and typological theory has never been clarified; indeed, the existence of typological theory is hardly recognized. Theories of grammar define representational formats and their interfaces. They determine what is describable (hence possible) and how this is learnable and computable. Distributional theories describe the statistical distribution of linguistic phenomena in time and space as well as in grammars (e.g. implicational hierarchies) and minds (e.g., Whorfian effects), and explain the functional, cognitive, and historical sources of this distribution. Typological theory defines the range of crosslinguistically viable notions that are assumed in both formal and distributional theories.

We have created a database of morphosyntactic information that applies (and even generates) pure typological information to create pure typological entries that are ready to serve as input to both formal and distributional theories. The status of the word is an issue that has been of concern to formal, distributional, and typological theories alike. We give an improved typological definition that clarifies problem cases of various kinds of isolating characteristics withhin synthetic words (e.g. Lai Chin, in which phonological bracketing tears apart richly composed polysynthetic words). A preliminary survey suggests that small phonological words (isolating languages) frequently go together with long grammatical words (synthetic to polysynthetic). We will use the example of wordhood to argue that formal and distributional theory are separately based on typological theory and that any feedback and explanation between formal and distributional theory is mediated by typological theory.

Joan Bresnan
Typology in Variation: An Stochastic Optimality Theoretic Approach
Empirical findings: Within a single grammar, variable outputs for the same input often reflect categorical generalizations of morphosyntax across grammars.

Example 1: The person hierarchy affects subject selection categorically in direct/inverse and active/passive systems (Silverstein, Aissen 1999).  It also affects the frequency of subject selection in active/passive choices in spoken English (Bresnan, Dingare, and Manning 2001).

Example 2: Individual patterns of variable subject-verb agreement with \be\ extracted from the Survey of English Dialects (Orton et al. 1962-71) show striking structural resemblances to grammars of categorical varieties (Bresnan and Deo 2001).

Theoretical explanation: In the framework of stochastic OT, it is not at all surprising that the variable outputs of individual speakers are constrained by the same kinds of typological and markedness generalizations that are observed crosslinguistically.  Typological variation across languages is explained within OT by means of constraint reranking.  In stochastic OT the noisy evaluation of candidates in individual stochastic grammars reranks these same constraints by temporarily perturbing their ranking values along the continuous scale.  As a result, individual variation samples the typological space of possible grammars.

Andrew Garrett
Synchronic vs. diachronic explanation of typological patterning
Numerous interesting generalizations have emerged from qualitative typological research in several areas of linguistics, including semantics (e.g. which kinds of verbs develop future meanings), syntax (e.g. which factors condition split ergativity), morphology (e.g. the prevalence of -VC- vs. -CV- infixation after initial C clusters), and phonology (e.g. which segment types undergo long-distance metathesis). I will contrast two approaches to explaining such generalizations -- (1) the relevant underlying principles form part of the linguistic or other cognitive system (so e.g. Aissen, Bresnan, Steriade, Sweetser); (2) typologically widespread patterns are epiphenomena of "convergent evolution" caused by contextual ambiguities in speech (so e.g. Blevins, Ohala) -- and I will give some arguments in favor of approach (2).

Larry M. Hyman
Compositional and/or templatic morphology: Suffix ordering in Bantu
In this paper I address the issue of templatic vs. compositional determination of affix ordering. Rather than constituting two distinct, i.e. incompatible types, I show that both determinants can in fact co-occur in the same language. Proponents of the "mirror principle" (Baker 1985), which claims an identity between the order of syntactic and morphological operations, as reflected in the order of affixation, find support in cases of apparent compositionality in Bantu verb suffixation, and proclaim the Mirror Principle to be "an exceptionless generalization" (Alsina 1999:6). However, in this paper I present several arguments that there is a default order of Bantu verb suffixes, schematized in the partial template:

Causative-Applicative-Reciprocal-Passive (CARP)

e.g. Chichewa -its-ir-an-idw-

I show that there is a "tension" between this template and compositionality (whereby an outer suffix should have scope over an inner suffix). I propose that language-specific overrides are responsible for licensing atemplatic suffix+suffix combinations, when they occur. This conflict is aptly captured via two families of constraints: TEMPLATE constraints and MIRROR constraints. I show that these constraints can have different rankings with respect to different suffixes and in different languages. Importantly, when they override TEMPLATE constraints, MIRROR constraints typically have a much more restricted and often idiosyncratic character. Since affix ordering is not directly predictable by appealing to the semantics (e.g. scope; Bybee's 1985 "relevance") or the syntax (e.g. Baker's Mirror Principle), this argues not only for the reality of templates in morphology--but for a morphology that is autonomous. The potential low ranking of a (syntactic OR semantic) Mirror Principle as a predictor of suffix ordering in Bantu shows that it is not universal in the "no exceptions" sense, but rather in the (violable) OT sense.

Elena Maslova
On linguistic interpretation of quantitative cross-linguistic data
Joseph Greenberg once wrote that, ideally, we must wish to assess probabilities of various typological shifts, but we are not in a position to do so. It can be easily observed, however, that any linguistic interpretation of a cross-linguistic statistical tendency implies certain (rather strong) assumptions on such probabilities. This fact is not always recognized, partly because quantitative cross-linguistic generalizations, in contrast with absolute implicational universals, are not easy to "dynamicize", i.e. to interpret as a set of specific constraints on language change. A general algorithm of such interpretation would provide a statistically reliable way to distinguish between genuine (linguistically motivated) distributional universals and "accidental" properties of the language population and, by the same token, generate a variety of testable hypotheses on the large-scale process of language change. The talk explores the possibility to develop such an algorithm. As a test example, I use one quantitative typological parameter, the number of vowels per language. The analysis is based on the data from (Maddieson 1984), supplemented with Ethnologue classification of languages.

Masayoshi Shibatani
The causative continuum
Traditionally causative constructions have been dealt with in terms of a formal trichotomy of (i) lexical, (ii) morphological, and (iii) syntactic or periphrastic forms (Comrie 1981), and in terms of a semantic dichotomy of (a) direct/contact/manipulative and (b) indirect/distance/directive causation (Nedjalkov & Sil'nickij 1969, Shibatani 1973). This presentation first points out that each of these two dimensions in fact forms a continuum and then establishes the significance of an intermediate causative type called 'sociative causation'.

Dan Slobin
Typological differences in representation of motion events: Some cognitive and communicative consequences
In order to produce or understand a linguistic message--in speech, sign, or written form--it is necessary to represent the content in terms that fit the lexical and morphosyntactic patterns of the particular language being used. This concept-to-language fit has consequences for how language producers attend to event dimensions, as well as for the imagined and remembered events that remain in the minds of receivers. Using motion events as an example, it will be suggested that the differences in lexicalization patterns between verb-framed and satellite-framed languages (Talmy) has widespread effects on discourse and associated cognitive processes.