Statement of Accomplishments and Objectives: 2004

Research

My research program is focused on the lexicon and its relationship to other major components of the language faculty. The lexicon is the locus of the most direct interface between language and thought. Understanding its nature is the underlying theme that unifies the disparate threads of my creative research.
Most current theories of lexical semantics and morphology (Pinker 1989; Bierwisch and Schreuder 1992; Pustejovsky 1995; Jackendoff 1997; Levin and Hovav 1998) and of morphology (Anderson 1992; Aronoff 1994; Beard 1998) assume that there is a generative component within the lexicon proper, which takes lexical elements (semantic and/or morphological) as input and generates more lexical elements, enlarging the lexicon. This lexicon then provides the input to the syntax. In such theories, language gets its fabled generative capacity from more than one source: a pre-lexical semantic/morphological generator and a post-lexical syntactic generator. To boot, linking rules associating semantic elements with syntactic positions are also necessary.
A competing hypothesis that has emerged in several forms over the past decade suggests that this division of labor is incorrect (see, e.g., Hale and Keyser 1991; Noyer 1992; Kratzer 1993, among many others). In particular, the framework of Distributed Morphology (Halle and Marantz 1993) suggests that generative capacity is solely the province of the structure-building syntactic component. That is, there is no separate word-building morphological component, nor is there a meaning-altering lexical semantic component, nor is linking theory necessary. Syntax combines abstract features, constrained by internal principles, and lexical items are then inserted to Ôrealize' these abstract features. Interpretation is compositionally read off the completed syntactic tree, in consultation with the Encyclopedia (the list of sound-meaning pairings). Lexical semantic phenomena, particularly argument structure alternations, are the product of the syntax.
I investigate these questions within several subfields of linguistics. Most of my work is aimed at discovering the nature of the basic elements of the grammar. The ultimate goal is to determine whether a syntax-based account of lexical generative capacity can provide gains in theoretical parsimony and explanatory power.

The Syntax of Lexical Semantics
In theories of lexical semantics with a generative lexicon (see citations above), the relationship between two words with a common root (for instance, intransitive break vs. transitive break, or the noun shelf and the verb shelve) is usually explained in terms of a lexical rule which manipulates the semantic representation of one form to produce the new semantic representation of the other. Only certain types of semantic element may be manipulated, however. In work in my dissertation (Harley 1995), I argued in favor of a treatment of the inchoative-causative alternation (break(tr.) vs. break(intr.), melt(tr.) vs.melt(intr.) etc.) in purely syntactic terms, drawing on evidence from Japanese. I proposed a return to a sharply constrained version of the decompositional approach of the Generative Semantics school of the early 1970s, limiting the extent of the syntactic decomposition by tying it to a single semantic event; this work appeared as Harley 1995; Harley 1996.
A consequence of that proposal was that all agentive verbs should decompose in the syntax into a Ôlight verb' (CAUSE) and a root, containing lexical meaning. Focusing on the root in the case of one specific class of verbs, those which undergo dative shift, led me to propose the existence of an underlying HAVE predicate in their syntactic representation (that is, Ôgive' really is, syntactically, CAUSE to HAVE. This work was first presented at conferences (Harley 1996), and later appeared in a journal (Harley 1997). It led to a more general investigation of the predicate have in English, which can take a dizzying array of syntactic complements, with a correspondingly wide variety of semantic interpretations. It turns out that its meaning is entirely dependent upon the syntactic complements it takes, lending support to the notion that there is only one lexical item have. This work led to two journal publications (Harley 1997, 2002) as well as a distinct article in a book (Harley 1998). I continue to pursue this line of investigation, reconsidering old arguments from the literature (Hall 1974, McCawley 1974, Dowty 1979) concerning the treatment of want with a noun phrase object (e.g. want a beer) as underlyingly want to have a beer. An article on this topic has appeared in a journal (Harley 2004).
I subsequently investigated the contributions of lexical roots that are shared by related words in the Hale and Keyser framework for argument structure. I found that denominal location/locatum verb roots (e.g., to shelve, to water) are marked for boundedness. When roots are realized in a nominal context (a shelf, water), this marking is realized as the mass/count distinction; when in a verbal context, it becomes the inherent telicity of the verb form. I presented these results at the Israel Science Foundation Workshop on the Syntax of Aspect at Tel-Aviv University, Israel, in January 2001, as well as at an invited colloquium talk at the University of Maryland in December 2001. It will appear as a chapter in an edited volume (Harley forthcoming, 2004). This line of research has also led to a jointly authored journal article on the remarkable syntax and semantics of Persian complex predicates with Simin Karimi and Raffaella Folli.
Since arriving at the University of Arizona, I have become interested in the semantics and syntax of verb classes in Hiaki, an Uto-Aztecan language spoken in southern Arizona. During the year 2000, I and Maria Amarillas, a Hiaki language teacher, have conducted a pilot study on verb classes and reduplication in Hiaki funded by the Office of the Vice President for Research at U of A. Hiaki has at least 3 different forms of reduplication. That research led to a grant from a foundation in Washington, the Jacobs Research Fund, and an article in a book on Uto-Aztecan in the MITWPL Endangered and Less Familiar Languages series, as well as a conference presentation this year.
I began another investigation of the syntactic representation of telicity with Raffaella Folli when she was a postdoctoral fellow at the University of Arizona, comparing telicity in English and Italian verbs of consumption. In both languages, telicity proved to be intimately connected to whether or not the consumer subject of the verb was animate or inanimate (compare John ate the apple (up) with The washing machine chewed the laundry *(up)). Our work on that topic was presented at the NSF-funded workshop on aspect held at the University of Iowa in May, 2002, and will appear in an edited Kluwer volume next year. Raffi and I have now begun to analyze the different kinds of Italian causative constructions, a line of work which has led to several conference presentations and a working-volume paper; we plan to submit a fully developed paper to a refereed journal later this year. Our collaboration has led to a grant from the British Academy for international collaboration, which will fund travel to Cambridge for me in the summers of 2004 and 2005; we hope to prepare a book manuscript and organize a workshop on telicity.
Work on the nature of v leads one naturally to a consideration of head-to-head movement in the Minimalist Program, since v is nearly inevitably realized in concert with a root, not independently. Investigation of this idea led me expand on a conflation mechanism first proposed in Hale and Keyser 2002. I observe that this mechanism could account for head-to-head movement and the observed constraints upon it, providing a theory of head-movement which meets the Minimalist Program requirement of being phonological, not syntactic, thereby resolving a fundamental tension in current syntactic theory. I presented that work at a symposium at the University of Maryland in May 2002 and at NELS 2003. A preliminary article reporting those results is appearing in the 2003 NELS proceedings, and fuller working-out of the idea is in progress.
Finally, in 2000 I began a Ôstraight' syntax project, investigating the relationship between weak crossover, covert quantifier raising and antecedent contained deletion. The resulting article appeared in a journal in 2002.

Morphology
No investigation of the lexicon could proceed adequately without raising many morphological questions. Is morphology an independent module, or can syntactic processes account for the word-formation as well as sentence-formation? How does the apparent lack of compositionality in the lexicon affect such a proposal? In work with Rolf Noyer of the University of Pennsylvania, I investigated constructions at the syntax/ morphology interface: verb-particle constructions and nominalizations. We showed that the variations in verb-particle constructions can be accounted for nicely by syntactic object shift (Harley and Noyer 1998). More interestingly, building on Marantz 1997, we provided additional evidence that some of the properties of nominalizations can be accounted for by Encyclopedic, that is, non-formal, information--real-world knowledge (Harley and Noyer 1998). Noyer and I also wrote the only extant comprehensive overview of the positions and results of the Distributed Morphology framework, Harley and Noyer 1999, 2002. That publication led to an invitation to teach a two-week intensive course on Distributed Morphology at the ABRALIN Linguistic Institute at the Universidade Federal de Rio, Rio di Janeiro, Brazil, in March of 2003.
In the DM framework, the syntax takes as input a collection of abstract morphosyntactic features. The question of exactly what these features are, and how they are organized overlaps in part with the questions of more traditional morphological theories. Using evidence from a large database of pronominal systems, Elizabeth Ritter of the University of Calgary and I proposed that person, number and gender/class features are subject to the same organization as phonological features: a feature geometry. We use the feature geometry to account for many otherwise mysterious generalizations about variations in, and the acquisition of, pronominal paradigms. In March 2000 I received an International Travel Grant to travel to Germany to present some of our findings at an international meeting (Harley and Ritter 2000) and in May I received an SBSRI Mini-Grant to travel to Calgary and collaborate with Ritter on revisions to our manuscript. That collaboration resulted in an article that appeared in a journal in the fall of 2002. I have subsequently been invited to participate in an international symposium on phi-features at McGill University in August 2004, funded by the Social Sciences and Humanities Research Council of Canada.
Finally, I have also investigated a relatively new word-formation process in English: the mophosyntax of acronyms and abbreviations. I observed that acronyms behave like proper names, syntactically speaking, while abbreviations behave like noun phrases. An article reporting this result will appear in a journal this year.

Psycholinguistics
The overarching question that drives my research abuts on many issues of importance to disparate fields of inquiry, and as such are amenable to psycholinguistic investigation as well. My appointment as a research professor in the cognitive science program has enabled me to take advantage of the expertise available in these fields at U of A, and to share my own.
I initiated an independent research project, with funding from the SBSRI Small Grants program, to investigate the factors that influence a speaker's choice of referent when they are exposed to a new word in a context where several alternative meanings are possible. Results indicate that very subtle differences in wording and presentation can make a significant difference in the meaning that speakers choose; the results are being written up for submission to Cognition. My results suggest that earlier work by others (Imai and Genter 1997), taken to indicate that the language a person speaks is a primary influencing factor in such decisions, is deeply flawed.
Another ongoing collaboration with Erin O'Bryan, Tom Bever and Raffaella Folli investigates the psycholinguistic reality of syntactic telicity. If telicity is indeed syntactically represented, it should have a strong effect on the processing of reduced relative clauses like The cart pushed by Bill (atelic) compared to The cart upset by Bill (telic). Previous investigations had considered the effect of transitivity on such processing. When we looked for the effects of telicity, we found that the effect of telicity was more than double the effect of transitivity, strong support from an unexpected quarter for the syntactic representation of event structure. Those results have been presented at several conferences and will be submitted to a journal this year.

Teaching

My overall goal as a teacher is to create courses that are both challenging and stimulating for my students, that leave them feeling that they have come to understand something new and interesting. I try to accomplish this in several ways: by incorporating recent research results into my class material, by approaching students' questions with an open mind, and by fostering independent discussion and analysis as much as possible. In my first years as a new teacher, I spent a great deal of time preparing course materials, and while my classes were successful, they sometimes lacked the extra creative touch that might have made them truly distinctive. Since then, however, I feel I have succeeded at least a few times at providing the key supplemental material or presentation of a concept that made the difference between a good and a great course for at least some students. I have taught many different courses since arriving at the U of A, almost never the same one twice. This has been extremely stimulating and educational for me, but as I come to repeat a few courses, I anticipate that I will be able to step back and assess the overall learning experience of my students with increasing objectivity. My student evaluation scores have tended to become stronger over the past three years, and I intend to see the trend continues.

Undergraduate
In teaching undergraduates, I have two main goals. First, I wish to send from the classroom a student with a new appreciation and understanding of their complex, unconscious, linguistic knowledge. My areas of expertise touch on many issues that are of interest to the educated layman, and about which there are many misconceptions. Getting students to understand the richness and complexity of human language is important, not only for my gratification as a teacher and for the intellectual interest that is inherent in the discipline, but also as a step away from the linguistic jingoism that so often goes hand in hand with more insidious forms of intolerant behavior. This is one domain in which a linguistic education can dispel commonly held, untrue and harmful beliefs.
My second goal in teaching undergraduates is to help students improve their analytical skills. Linguists are pattern-finding, puzzle-enjoying types, and linguistics majors are often attracted to the field by this aspect of it. Students who are not natural puzzle-solvers, however, need to be introduced slowly and thoroughly to the analysis and synthesis necessary to solve linguistic problems. Such an introduction, of course, can help students with problem solving in many domains. A linguistics class is an ideal place to inculcate and nurture analytic skills, and it is important to me to exploit that opportunity. In the same vein, I wish to help students become capable independent researchers, with a knowledge of how to use library databases and journal articles to answer questions in linguistics and other fields.
Both of the undergraduate courses I commonly teach, Linguistics 322 and Linguistics 211, have no prerequisites other than an INDV course. They attract a large number of students, many of whom are taking their first linguistics class. This can create a difficult situation, where the linguistics majors are frustrated by necessary repetition of material they have learned elsewhere, and non-linguistics majors find themselves somewhat bewildered by the many new concepts they are required to assimilate. I have attempted to address this problem in several ways. I have had linguistics majors volunteer as tutors for non-majors, with the intention that the linguistics majors acquire a better grasp of key concepts through helping teach them to the non-majors. I have also allowed the linguistics majors to test out of certain basic section of 322 (e.g. learning IPA transcription). Despite the mixed audiences, these courses have became reasonably successful. Teaching 322 has also inspired me to write a textbook, to appear with Blackwell ("English Words: A linguistic introduction"). In the textbook and in class, I find it is particularly important to bring in recent psycholinguistic results concerning the mental representation of the lexicon, and that forms a major component of the discussion in the textbook as well. Teaching 322 inspired me to begin the psycholinguistic investigation of word learning discussed above, and in both 322 and 211 I have presented the results of this research. In 211 I have incorporated a few in-class versions of others' psycholinguistic experiments using Powerpoint technology; this has been one of my most successful teaching strategies. I also have used multimedia techniques in both courses, playing clips of experiments, music and readings in Old English, which I find can help to engage the students well.
I often supervise independent studies courses for undergraduates. Two in particular are worth commenting upon in that they directly involved undergraduates in my research activities. Undergraduate Jeffrey Watson worked with me on a project to survey the grammatical properties of VSO languages. That project is part of a larger research project with Andrew Carnie and Sheila Dooley Colbert which has recently received an NSF grant to sponsor a workshop on VSO languages. Maria Amarillas is another undergraduate who enrolled in two independent studies courses with me. She is a Yaqui (Hiaki) language teacher and religious leader. I worked with her to apply for the McNair Achievement Program, and acted as her faculty mentor during her participation in the program. She is the primary consultant for the study of Hiaki verb reduplication now underway; her contribution is of course invaluable; the project could not proceed without her.

Graduate
At the graduate level, I view my role more as a facilitator and mentor than strictly as instructor. Ultimately, we wish to produce independent professional researchers, with good research and teaching skills and a coherent research program. It is important in classes to foster a collegial atmosphere, encouraging students to produce original work related to the themes of the course.
I have taught more graduate than undergraduate courses since arriving at Arizona, and consequently I have interacted with the majority of our current graduate students. I mostly teach from original literature, sometimes assigning a textbook as backup reading in introductory courses in syntax and morphology. My work with our graduate students is one of the most rewarding aspects of my job at the University of Arizona, and I place great value on the chance to contribute to their education and original research. I currently supervise seven dissertating students; the first dissertation defense of a student I've supervised is scheduled for July 2004. My seven dissertators are working on the syntax and morphology of Uto-Aztecan, the null pronouns of Brazilian Portuguese, the morphology and semantics of the subjunctive in Slavic and Romance, the syntax of Pima Bajo, the syntax of Russian, the morphosyntax of Persian, and the morphosyntax of Korean. It is wonderful for me to get to learn more about the properties of all of these languages; I find that supervising is at least as enriching for me as it is for them.
I am a faculty member in the unique Anthropology/Linguistics (AnLi) program offered at the U of A; in that capacity I have served on AnLi students' preliminary exam committees, and advised them on their research program. Two of my dissertating advisees are AnLi students.
Finally, I have assisted Mary Willie and Ofelia Zepada with the M.A. program in Native American Linguistics by giving a few guest lectures. I hope to become more involved with the program, including by mentoring to students (like Maria) who can benefit from one-on-one attention.

Service

I have participated a fair amount in departmental activities, but thanks in part to the policy of sheltering junior faculty from overmuch committee service, I have not served on many college-wide or university-wide committees. Within the department, I have served on several committees, including hiring, head search, outcomes assessment, departmental self-study, colloquium, and peer-review committees. I also consider it very important that the department have a voice when relevant local political or social issues arise, and in that connection I drafted a departmental position paper on Prop 201 in the fall of 2000 and attended press conferences and rallies to articulate departmental views. In a similar vein, as part of an effort to contribute something substantive to the Yaqui tribe who permitted me access to their language, I constructed a picture book with a Yaqui children's story for use with the small children in the tribal Head Start program.
Many linguistics journals, conferences, and publishers have requested my services as a reviewer, which I generally accept as a great chance to stay acquainted with cutting-edge research in my areas of interest. I also helped to coordinate the visits to campus of Beth Levin and Howard Lasnik. In general, I view service activities as an opportunity to expand and improve the reputation and influence of the department and institution with which I am affiliated.