Click Here! For PDF Version of the 2011 Colloquium Series
The University of Arizona Department of Linguistics
Friday 3pm-4.30pm in Communication 311
Coordinator and convener: Massimo Piattelli-Palmarini (Linguistics, Cognitive Science and Psychology, University of Arizona)
Friday, August 26th
Sandiway Fong (Department of Linguistics and Department of Computer Science, the University of Arizona)
Title: On the generative capacity of natural languages: human and non-human
There has been recent interest on distinguishing the generative capacity required for human language versus language for other animals. There has been a long history of discourse on the generative capacity of human language. For example, results involving center-embedding phenomena and cross-serial dependencies point to the fact that human language is intrinsically non-finite state and non-context-free, respectively. There has been further debate as to whether human language belongs to the class of mildly context-sensitive languages. However, given our inability to properly deconstruct utterance meanings for non-humans, there have been comparatively few results about the generative capacity of animal utterances. Nevertheless, there are observable constraints about the surface structure of non-human utterances. For example, birdsong has been analyzed as being of finite state capacity. In this talk, we will analyze the utterances of non-human primates from a formal language perspective. In particular, we will suggest that northern muriquis, also known as woolly spider monkeys, generate utterances that are non-context-free: in fact, beyond mildly context-sensitive in nature. (Joint work with Didier Demolin, GIPSA Lab, Grenoble, France.)
Friday, September 2nd
Orientation for first and second year students
Graduate Advisor Dr. Diana Archangeli and Graduate Program Coordinator Jennifer Columbus will define Linguistics graduate programs' requirements/expectations. Other UA Administrators will also attend, as well as Linguistics Department Program Faculty. All incoming graduate students, and 2nd year graduate students in NAMA, HLT, ANLI & LING PHD are required to attend the Department Orientation.
Friday, September 9th
Jeffrey Punske (Department of Linguistics, University of Arizona)
Title: Morphological Conspirators: morpho-syntax and the apparent event-structure differences across forms of
Morphological Conspirators: morpho-syntax and the apparent event-structure differences across forms of nominalization Sichel (2010) noticed a difference in the types of events permitted in nominal gerunds and derived nominals: that nominal gerunds allow particle constructions and resultatives while derived nominals do not.Sichel correctly notes that complex verbal expression (resultatives, particles) are permitted in nominal gerunds but not in derived nominal (below, modified from Sichel 2010).
1. *John’s explanation (away) of the problem (away)
2. John’s explaining (away) of the problem *(away)
This generalization runs counter to Harley’s (2005) account of Latinate bipartite verbs which Harley argues are structurally identical to particle constructions since Latinates heavily populate the derived nominal construction. Sichel’s account faces an independent problem in predicting that derived nominals should be syntactically less complex than nominal gerunds (following Travis 1992, Borer 2005, Ramchand 2009) which has no independent support beyond the missing constructions and may be demonstrably incorrect if Punske’s (2010) account of nominal argument structure is correct. Given these problems, I offer an alternative account of these facts which does not require derived nominals to be syntactically less complex; instead, I show that the apparent restriction is actually a conspiracy of morphological selection restrictions that creates the appearance of an event-structure restriction. Without reference to event structure the apparent restrictions on complex events in derived nominals can thus be explained. Derived nominals are morphologically ill equipped for verb particle constructions because particle incorporation interferes with the syntactic merger of the root and n0. Nominal gerunds are not burdened by the requirement of syntactic merger of the root and n0 (merger happens post-syntactically) so the root can incorporate without interfering morphologically. Similarly, Latinate forms generally disallow particles, so their failure to allow particles in derived nominals is anything but surprising. Since Latinate verbs are complex events themselves, the event analysis offered by Sichel cannot be maintained.
Friday, September 16th
Mariette Marsh (Assistant Director, Human Subjects Protection Program, The University of Arizona).
Recent changes in the Human Subjects Protection Program
The colloquium will focus on the newly released changes to the forms and policies by the Human Subjects Protection Program effective August 1, 2011. All investigators should attend this important session. Time permitting, discussion may include the proposed new rulemaking concerning the human subjects regulations.
Friday, September 23rd
(in collaboration with the Department of Philosophy)
Robert C. May (Department of Philosophy and Department of Linguistics, University of California, Davis)
Title: Leibniz's Problem, Frege's Puzzle: Language and representation of scientific content
In this paper, I will present a portrait of Frege as centrally concerned to make precise how scientific knowledge can be derived and expressed. At the core of this enterprise are Frege’s central linguistic innovations as these are embedded in the Begriffschrift, his conceptual notion. In this context, an issue appears to emerge when Frege applies logic to scientific subject-matter, starting with arithmetic: how is language representationally connected to content qua subject-matter?
This problem is Leibniz’s Problem: it manifests itself as a problem about identity: Why doesn’t ‘2+3 = 5’ reduce to ‘5 = 5’ and thus express no mathematical content above and beyond the trivial? What I will propose is that Frege does not so much address this issue but reconceptualize it as Frege’s Puzzle. He makes out the problem of identity as not so much as about propositional content as about our cognitive relation to ir, and how this gives rise to scientific knowledge. In Frege’s mature conceptual framework, the lesson of hte puzzle of identity is that scientific inquiry is sensitive to the cognitive value of propositions (thoughts). Central to Frege’s approach, I will argue, is his elucidation of the representational role of the Begriffsschrift through the semantic and cognitive roles played by the sense/reference distinction in his account of judgement
Friday, September 30th
Tim Stowell (Department of Linguistics, University of California, Los Angeles)
Title: Elliptical adjuncts
My talk is concerned primarily with the double adjunct (DA) construction in (1):
1 a. The rebels have been defeated, perhaps decisively.
b. Jack drank an entire bottle of gin last night, reportedly in ten minutes.
In 1a, the adjunct sequence consists of two adverbs: perhaps (A1) and decisively (A2); in 1b, A1 is again an adverb (reportedly), and A2 is an adjunct PP conveying duration. In DA constructions more generally, A1 is typically a ‘higher’ adverb (evaluative, evidential, modal, etc.) and A2 is typically either a ‘lower’ adverb or almost any other type of adjunct that can occur independently in post-verbal position.
DA constructions, like adverbs in general, can occur sentence-initially, sentence-finally, and sentence-medially—generally, somewhere between the subject and the main verb, either before or after various combinations of auxiliary verbs.
2 a. The rebels have been defeated, perhaps decisively.
b. The rebels have, perhaps decisively, been defeated.
c. Perhaps decisively, the rebels have been defeated.
The DA sequence as a whole is segregated intonationally from the rest of the sentence by what is traditionally called comma intonation. In particular, the DA sequence is separated from the rest of the sentence by a pause and it forms an independent intonational phrase. Comma intonation is accompanied here by parenthetical semantics. The DA sequence is not integrated semantically into the primary assertion of the assertion, which is just the content of the main (host) clause. The DA is interpreted as conveying a secondary assertion, with a discourse function that has sometimes been characterized as a comment on the main clause.
The propositional content of the secondary assertion provided by the DA informs the syntactic analysis. I will argue for an analysis of the DA construction based TP ellipsis, namely the same syntactic mechanism that Jason Merchant has employed to account for the derivation of Sluicing constructions. More precisely, the DA sequence in 1a is derived by TP ellipsis from a source structure resembling 3a, perhaps involving an intermediate structure 3b, with ellipsis deriving either 3c or 3d. The target of ellipsis is identical to the main clause hosting the DA; thus, the main clause is the antecedent of the elided TP. (Note that, at least in the case of 2b, the application of ellipsis may involve a type of ACD.)
3 a. Perhaps the rebels have been defeated decisively
b. Perhaps [decisively] the rebels have been defeated
the rebels have been defeated decisively
d. Perhaps [decisively]
the rebels have been defeated
Other types of DA constructions will be explored in some detail, including: constructions where the linear order of A1 and A2 is reversed; structurally complex constructions where A1 is replaced by two or more constituents; and other variants where A2 is replaced by an argument. I therefore suggest a more general name for the construction, namely Elliptical Adjunct (EA) constructions. The variation is predicted naturally by the ellipsis account, given the existence of plausible source structures for every type of variant. Furthermore, the parallel with Sluicing is strengthened by the examples where A2 is an argument.
These parenthetical DA constructions will then be compared to simple adverbs set off by comma intonation, with respect to their syntactic distribution and parenthetical (or non-parenthetical) semantics, to determine (a) whether the ellipsis analysis carries over to these less complex structures, and (b) to what extent the ellipsis analysis contributes to our understanding of the phenomenon of variable adverb placement, of the sort illustrated by the DAs in 2, with and without comma intonation.
Friday, October 7th
Robert Mailhammer (Department of English, Arizona State University)
Title: Regularity as a continuum: why “irregular” doesn’t have to be the opposite of “regular”
In linguistics, especially in morphology, the word regular is typically used in an extension of its everyday meaning, which is ‘following a pattern, rule’ (and therefore predictable see e.g. OED, s.v. regular), i.e. the opposite of e.g. ‘erratic, unpredictable’. As a result, it chiefly means ‘usual, normal’, and therefore the term regular verb conjugation in Modern English, Latin, French etc. refers to the inflectional pattern that the majority of verbs follow. However, this is not ideal from the viewpoint of a parsimonious taxonomy. If regular, actually means ‘normal’, i.e. ‘what the majority does’, then what do we call something that follows a pattern but just not that of the majority?
This problem can be illustrated with examples from Modern English: While play, dance and study are called ‘regular’, sing, sink, drink and ring are called ‘irregular’, alongside cases such as come, be, catch, sleep and go. It is, however, apparent that sing, sink, drink and ring follow a pattern, which has even been shown to be productive for nonce words (Bybee & Slobin 1982). Moreover, the inflectional pattern of sleep is shared by other verbs, such as leap and sweep; catch is a loanword that nevertheless belongs to an ‘irregular’ inflectional type. Verbs like come did actually once belong to an inflection class, whereas go and be are much more obviously irregular, as they show non-predictable suppletion. To call all these verbs irregular obscures these finer distinctions and their linguistic and cognitive implications. For instance, local, minimal generalisations based on phonological similarity have shown to be much more reliable, much more regular (in the more general sense of the word) than unspecified default rules that are typically called ‘regular’ (see Mailhammer 2007 with references).
This talk explores the idea that regularity is of much more use to linguistics if taken in the more everyday sense (‘following a pattern, rule’). Using examples from a variety of languages, I will posit that this kind of regularity actually represents a continuum stretching from strict uniformity to fairly erratic suppletion, deponency and defectiveness. I will outline fresh investigative perspectives that this perspective opens up, including quantification, psycholinguistics and diachrony (natural languages are remarkably irregular), as well as a typology of regularity.
Friday, October 14th
Terje Lohndal (Department of Linguistics, University of Maryland)
Title: Spell-Out, Movement, and the Copy Theory
This talk explores the consequences of a dynamic theory of Spell-Out where two phrases can never be merged. I will show how this theory provides a transparent mapping onto Neo-Davidsonian logical forms. Then I will scrutinize syntactic consequences of the proposal, involving the copy theory and movement. I will discuss how linearization works and how this theory can provide us with an account of the EPP in English, and I will discuss how it utilizes a conception of movement close to that of Nunes (2004). I will also show how the theory effectively bans multidominance structures.
Friday, October 21st
Norbert Hornstein (Department of Linguistics, University of Maryland)
Title: A Movement Theory of Pronominal Binding
This talk follows Kayne 2001 in modeling pronoun binding as a species of A to A’ to A movement, aka “improper movement.” The aim will be to demonstrate the theoretical and empirical utility of so treating pronominal binding. There are several theoretical upsides should this prove viable. At the most abstruse level, this would allow for the complete reduction of Binding Theory to movement theory, thereby advancing the minimalist aspiration of reducing all grammatical dependencies to species of Merge. Second, coupled with a movement theory of local anaphora (e,g, Hornstein 2001, Zwart 2002) it is possible to derive a simple explanation for the complementarity of bound pronouns and reflexives as well as the c-command condition on binding. Third, assuming pronominal binding to be the product of movement allows us to derive several conditions on bound pronouns (e.g. Strong and Weak Crossover effects, subject orientation effects) from more general conditions on structure building (viz. Extension and Economy).
Truth be told, these theoretical virtues are enough for me. However, there are apparently still linguists saddled with an obsessive prurient interest in the “data.” To humor you I will consider what I believe to be novel data showing the ECP inducing effects of pronoun binding, curious subject/object asymmetries with respect to the binding of definite descriptions, and apparent island sensitivity of pronominal binding.
If there is time (which I strongly doubt) I will try and contrast this approach to pronoun binding with the one I developed in (2007). The latter has two very interesting features. First, it derives the semantic inertness of bound phi-features. Second, it provides a nice (though most likely empirically incorrect) account of obviation effects in romance subjunctives. I believe that (at least) one version of the movement account can replicate these positive features and I would like to outline how this might be done in the context of a movement account of pronominal binding.
Friday, October 28th
3pm, Marshall Building, Room 490 (please notice the change of building and room)
Joint event: Prof. Marc Aronoff (Department of Linguistics, Stony Brook University) will give a talk as one of the guest speakers at Arizona Linguistics Circle 5. His talk will be on the Al-Sayyid Bedouin Sign Language and will be scheduled as a joint Arizona Linguistics Circle-School of Middle Eastern and North African Studies. His talk will be followed by a film screening event (a documentary on this recently discovered sign language in a totally isolated community of the Negev desert).
Title: Al-Sayyid: A Language Blooms in the Desert
Al-Sayyid Bedouin Sign Language (ABSL) was created about 75 years ago by the first deaf people born into the village of Al-Sayyid in the Negev desert, four siblings. Their deafness originated in a nonsyndromic recessive genetic trait that can be traced to the founders of the village and that is now widespread in Al-Sayyid. Today, there are close to 150 deaf members of the Al-Sayyid community, all of whom grew up with ABSL as their first language, and the language is used by many of the 3000 hearing members of the community as well. For the last decade, a team of linguists has been closely studying all aspects of the structure of ABSL at the level of syntax, prosody, phonology, morphology, and lexicon. This lecture will show how these distinct levels of structure have emerged in ABSL, each in a quite different fashion and at its own rate.
Friday, November 4th
James Higginbotham (Department of Linguistics and Department of Philosophy, University of Southern California)
Title: De Re and De Se: Linguistic and Conceptual Parameters
In recent years the issue of first-personal, or as David Lewis called them de se, thoughts and/or propositions has become further vexed, with several very different views and considerations in play. In this talk I will first briefly review a perspective on the subject that I advanced in the Sheffield meeting in 2000 (published in 2003, and available through Lingbuzz), and subsequently extended, most recently in a joint work with Marina Folescu (Cambridge: to appear). I then take on both some general semantic questions, and some linguistic parameters that appear to be at work in the de se, de re, and in other cases of embedded indexicality, for instance as examined in Pranav Anand's 2006 MIT dissertation. The topics will be interconnected, in that issues of the first person (and embedded indexicals generally) bind together questions about the nature of thoughts, propositions, and properties, and their representation through language, answers to each having consequences for the others.
Friday, November 11th
Veteran’s Day NO COLLOQUIUM
Friday, November 18th
Elly van Gelderen (Regents' Professor, College of Liberal Arts and Sciences, Arizona State University)
Title: The Diachrony of Pronouns and Demonstratives: Feature Economy meets Old Norse.
This talk provides a description and account of some of the changes involving the DP, namely anaphoric marking in English (and Scandinavian). I argue that Old English personal pronouns are not deictic/referential but that demonstrative pronouns have this function. This situation reverses itself in early Middle English, due to both internal and external factors. The internal factors are the shift towards the use of demonstratives before noun, i.e. the introduction of an article; the external factors are language contact that introduces new personal pronouns. I also look at Old Norse where the use of pronouns and demonstratives is similar to that in Old English. This remains the case to a lesser degree in its modern descendants.
Friday, December 2nd
Barry Schein (Department of Linguistics, University of Southern California)
Title: Frame, Count, Action. Semantics in Hollywood.
Frequent flyers recounted in their frequent flights under (1)’s protocol keep (1) from implying (2):
(1) Four million passengers crowded National Airlines’ routes last year.
(2) # Four million frequent flyers crowded National Airlines’ routes last year.
(after Gupta 1980: 23, Moore 1994).
If the arithmetic property four million is determined by the numeric identity of what is referred to, four million passengers does not refer to the one million persons who flew frequently:
(3) (4,000,000(X) & 1,000,000(Y)) → X ≠ Y
Yet, the identity of passengers and frequent flyers denied in order to thwart the inference from (1) to
(2) is asserted—truly, it seems— in (4):
(4) The four million passengers who crowded National Airlines’ routes last year were the one million frequent flyers loyal to it.
Worse, the four million passengers of (1) who are fewer persons do not always exist as such, subject to a condition that Doetjes & Honcoop (1997) call sequencing of events:
(5) Four million passengers had four million opinions about the food on National Airlines.
(6) # Four million passengers have four million opinions about the food on National Airlines.
Sentence (1) can be followed by (5), recounting the passengers’ experience onboard and continuing to count as many the fewer. In the present tense, however, (6) cannot count more passengers than persons who fly, despite the fact that these persons still have four million opinions about flights they remember all too well. Do they who once numbered four million no longer exist?
Instead of ‘four million(X)’, let four million translate as something like “now counted to four million”, ‘count[E,X, 4000000]’, a description of events of measurement (the counter clicks, imagine). What is counted and what the indefinite description four million passengers denotes is naïvely what there is, persons, who are counted and under some protocols recounted. If the logical form of sentences turns nominal descriptions into adverbs (adverbialization), with counting now part of the description, sentences (5) and (6) become (7) and (8) respectively:
(7) Four million passengers while counted to four million had four million opinions about the food on National Airlines.
(8) #Four million passengers while counted to four million have four million opinions about the food on National Airlines.
Clicking the counter as passengers go by frames or locates opinion recorded contemporaneously as in (5); but, there is no counting to four million that frames current conditions, which, according to (8), is what defeats (6). The speaker is secure in her constant reference to naïve, familiar objects by a sleight of hand that manipulates her epistemic conditions, in this case, conditions of measurement.
In turn, the translation of four million may be further refined to ‘count[E,X, F, 4000000]’ to include a parameter F for the frame or frames of reference under which the protocol for measurement is conducted. There is then daylight among (9)-(11), no two of which are synonymous, in that (9) and (11) engage distributive quantification over frames of reference; and, the contrast between many elms
in (12)-(13) and some/the many elms in (14)-(15) reduces to that of plural and singular reference to frames of reference, without syntactic or semantic ambiguity in the morpheme many:
(9) many an elm
(10) many elms
(11) many a one or more elms
(12) * Many elms are a cluster in the middle of the forest.
(13) * Many elms that are a cluster in the middle of the forest are diseased.
(14) The many elms are a cluster in the middle of the forest.
Some (very) many elms are a cluster in the middle of the forest.
(15) The many elms that are a cluster in the middle of the forest are diseased.
Some (very) many elms that are a cluster in the middle of the forest are diseased.