Year 1 activities and findings: Brief report
Research and education activities
The central goal of year 1, as described in the project description, was to gain a better understanding of the factual situation regarding expressives cross-linguistically. At the time, relatively little theoretical work had been conducted on expressives outside of English, and the focus even within English had been extremely narrow. So it was imperative to broaden the empirical domain. This was the primary objective, and the participants went after it wholeheartedly. There are relatively few researchers involved in the project, but they are expert in a variety of languages, especially English, German, Japanese, and Spanish. In addition, project participants collaborated with linguists working on Hindi, Navajo, and Tibetan. The result of this work is a large body of empirical evidence concerning expressives. This is, in and of itself, a valuable contribution, since these items are seldom discussed in descriptive grammars (because they are, by and large, taboo).
In July 2007, Potts taught a month-long course titled Dimensions of Meaning at the LSA Summer Institute, held at Stanford. Two of the lectures were devoted to the core phenomena of expressive content, and all of them were related in various ways to those issues. About 60 students and professors attended. They came from a wide variety of educational backgrounds — beginning undergraduates, advanced graduate students, international scholars from other fields, and tenured professors in linguistics. Potts therefore sought to find a balance between formal development and informal exposition. The lectures were data-driven, which provided students with a chance to offer evidence and insights from their own linguistic backgrounds. At the end of the course, students wrote short papers. Many students focused on expressives and, because the area is so new, they found a considerable amount of new data.
This project offers many chances for collaboration, both within linguistics and outside of it. One of the stated goals of the project was to capitalize on the ongoing NSF project in UMass Amherst Linguistics on the morphosemantics of evidential marking (Speas, Roeper, de Villiers, and Garfield; NSF grant #BCS-0527509). This collaboration has begun: Speas, Potts, and Christopher Davis collaborated on a conference presentation and an associated publication. This was Davis's first international conference, and it was the first time that Speas and Potts had collaborated formally. The paper concerns the pragmatics of evidentiality. It develops the idea that evidentials mark a particular discourse strategy, and it argues that a variety of different semantic treatments are compatible with the strategic approach. Since that publication, all three participants have capitalized on the ideas in new ways: Davis in his research on Japanese particles, Speas in her work on the relationship between modals, presuppositions, and evidentials, and Potts in his research on decision-theoretic pragmatics.
Throughout year 1, the project participants worked to achieve the goal of using expressives to inform questions about the general shape of linguistic theory. Two subquestions that received considerable attention were the nature of evidential marking and the importance of dynamic semantic interpretation. Evidentiality is the subject of the collaboration described just above, and Potts and Davis have been exploring the dynamic questions. Expressives pose special challenges in this area because of two generalizations about them that seem to be in tension: (i) expressives are interpreted as though they had widest possible scope; and (ii) expressives are interpreted at the point at which they sit in the utterance. In February 2008, Potts visited NYU to lecture in Chris Barker's seminar on dynamics and discuss these questions with Barker, who has worked on constructions that display a similar mix of properties.
The year 1 research goal — more evidence from more languages — involved a lot of library-time, so to speak, for the research assistants. Thus, it needed to be set alongside the goal of ensuring that those assistants presented and published their work, to further their professional training and to make the project's results more widely known. The project was successful in supporting its participants in this way.
As discussed above, the project participants devoted much of year 1 to broadening the empirical foundations of the project. The picture that is emerging from this work is a complex one. Expressives are morphologically, syntactically, and semantically diverse. Some are bound morphemes, and others are constructional. They are often functional, closed-class elements (as in Japanese honorifics), but there are open-class expressives as well — subclasses of common noun that allow for considerable inventiveness, for example. However, an abstract pattern emerges from a full survey of these data: expressives are, in an important sense, privileged. They are often freed from the normal morphological and syntactic restrictions of the language. This is a major lesson of the short collaborative paper Expressives and identity conditions (Linguistic Inquiry): in a wide variety of constructions that require matching between two elements, expressives constitute a systematic class of exceptions in that the are freed from the matching requirement. Similarly, project research assistant Florian Schwarz discovered that epithets do not participate in the process of prepositional incorporation normally found with definite nominals in Standard German. The facts are reminiscent of the special definiteness marking possible with epithets in Lebanese Arabic, and they arguably emerge subtly in English too, as a preference for demonstratives in epithets (e.g., That bastard Kresge). Project research assistant Christopher Davis has been investigating some productive bound expressive morphemes in Japanese, and they seem to permit what would otherwise be violations of restrictions on gemination and syllable final segments in that language. The facts indirectly recall the oddity of English that one can infix swear words (abso-friggin-lutely), but no other modifiers (*abso-stunning-lutely).
Thus, expressives can be privileged phonologically, morphologically, and syntactically. They are also subject to a host of their own special distributional restrictions. Thus, they seem to carve out their own linguistic space, which lends credence to the notion that they are linguistically special. This seems natural when set alongside the evidence, presented by Timothy Jay and Kristin Janschewitz in a 2007 published commentary on a paper by Potts, that expressives illicit different physiological responses than descriptive language. The findings could play a role in debates whether the language should be subject to special restrictions in public discourse.
All this new empirical data feeds directly into an overarching hypothesis of the project: expressives belong to their own dimension of meaning, separate from the usual descriptive semantic dimension. Defining and defending this thesis is the major goal of Potts's 2007 open peer-review article The expressive dimension, published in the journal Theoretical Linguistics with commentaries by Pranav Anand, Bart Geurts, Timothy Jay and Kristin Janschewitz, Peter Lasersohn, Philippe Schlenker, Uli Sauerland, and Malte Zimmermann. In that paper, Potts presents six criteria for expressive content. These evolved from the criteria presented in the description for this project, and they were informed by the data gathered during its early months.
The published commentaries on the paper do not dispute this multidimensionality, but they do question Potts's hypotheses about its sources. Three of them propose that expressives denote presupposition triggers, and thus that they fall together with that well-studied class of meanings. The hypothesis is initially appealing: it should mean that we can apply the rich set of tools and techniques from the theory of presuppositions to this domain, and that relatively little novel theoretic apparatus is required. However, this assimilation is untenable. In his reply to the commentaries, Potts shines a spotlight on the conceptual and empirical problems with it. However, he uses mostly familiar data to make the case. The above-described morpho-syntactic peculiarities were discovered only later, by project participants and their collaborators. Those peculiarities are simply not found with the usual class of presuppositional items. We can accept the label 'presupposition', but a great deal of new research will still have to be done to understand expressives in their own terms, and this will be a focus of the project over the next two years.
We have also addressed the combinatoric possibilities for expressives. In his earlier work, Potts took the view that expressives were rather limited in their modes of semantic composition. In late 2007, Davis began finding evidence expressives can combine in complex and productive ways with the material around them. His initial evidence was from a Japanese morpheme that productively maps descriptive adjectives into their expressive counterparts. He has since found preliminary evidence that their are other such morphemes. This is an important development because much of the early literature on expressives does not explore their distributional flexibility, either because that was not the researcher's focus or because the expressives under investigation were so distributionally limited as to be uninteresting in this regard. This project's work will show that expressives are integrated into the language, and, in doing so, it will highlight the diversity and complexity of the expressive realm.
It is emerging that we won't have a well-rounded picture of why expressives are special until we understand them as discourse moves. While we've seen that it can be fruitful to ask about the conventionalized content of expressive items, this might miss what's really special about these morphemes and constructions: they guide the pragmatic enrichment of the utterances that contain them. In choosing 'Sam's damn car is in the friggin driveway' over its expressive-free counterpart, I help convey to you how you should understand what I am saying and why it is important. Potts and Maria Biezma have begun exploring these issues — Biezma in her work on exclamatives, and Potts in his work on prenominal adjectives. In the coming year, this is likely to lead to the further integration of this research project with Potts's work on decision theoretic pragmatics and Davis, Potts, and Speas' work on evidential discourse strategies.