Workshop on Morphology and Argument Encoding

Workshop on Morphology and Argument Encoding
	Abstracts Back to Program Daniel Harbour: A Feature Calculus for Silverstein Hierarchies Handout as .pdf file Silverstein (1976) observed an important implicational relation between person and grammatical function. Consider the hierarchy below (a simplified version of Silverstein's): 1/2 > pronominal 3rd > human > other According to Silverstein, if, in a language L, [+human] NPs (say) receive special marking when the direct object of the verb, then all hierarchically higher NPs (in this case third person pronouns and local arguments) receive special marking when they too are direct objects in L. Conversely, if pronominal third person arguments (say) receive special marking when external arguments in L, then all hierarchically lower NPs (in this case, [+human] and other arguments) receive special marking when they too are external arguments. I propose an account of these facts that extends Adger and Harbour's (2007) analysis of syncretisms frequently concomitant with Person Case Constraint in unrelated languages. The core idea I will build on is that there are two different sources for phi-features on arguments in the syntax. On the one hand, heads minimally bear the phi-specification that their semantics demands. For instance, first and third person plural must both bear features for plurality, but only first person plurals must also bear a feature specification for person. On the other hand, heads may receive phi-features from the sister with which they are merged. Thus, third person arguments, for instance, differ in their phi-specification depending on whether they are merged with the verb or with a featurally contentful head such as v. Thus, an argument low on the Silverstein Hierarchy may, in virtue of the head that it merges with, receive a greater phi-specification than it intrinsically needs, and, conversely, an argument high on the hierarchy may endow the head with which it merges with more phi-features than it, the non-argument, intrinsically bears. On this conception, the special marking observed by Silverstein receives a straightforward analysis: it is the realization of the non-intrinsic phi-features that heads comes to bear in virtue of relations established by initial merge. The implicational relations (e.g., if [+human] NPs are morphologically marked as objects, so are all hierarchically higher NPs) follows from entailment relations between feature specifications. Petr Biskup: Prefixes and T-Features Abstract as .pdf file Handout as .pdf file Hakyung Jung: Genitive Subjects and Ergativity in North Russian It has been argued that ergativity is syntactically reduced to a structure in which the case-marking of the external argument is ascribed to a functional node lower than the matrix tense, which permits the internal argument to check the Nom against the T or the Acc against v (Woolford 1997, Ura 2000, Anand & Nevins 2006 among others). In this respect the so-called impersonal passive in North Russian with a genitive subject and a nominative/accusative object resembles an ergative construction. This paper examines the underlying structure of the given construction and evaluates it adequately in light of a typological perspective. I suggest that ergativity is not confined to a particular case morphology but appears as a general case marking strategy tied to certain syntactic positions in many languages, where an oblique case(s) is utilized to mark a subject. Gereon Müller: On Deriving Paradigm Economy Effects It is shown that assuming instances of syncretism with argument encoding to be systematic in the unmarked case may significantly reduce the number of possible inflection classes that can be generated on the basis of a given inventory of markers, without recourse to specific constraints like Andrew Carstairs-McCarthy's Paradigm Economy Principle or No Blur Principle. If there is always one radically underspecified (i.e., elsewhere) marker per morphological domain, and if there is always one unique marker that is chosen in cases of marker competition, it turns out that there can be at most 2^n-1 inflection classes for n markers, independently of the number of instantiations of the grammatical category that the markers have to distribute over. The argument relies on the notion of marker deactivation combinations. Cedric Boeckx: Isolating Agree In this talk I examine various instances of long-distance agreement (LDA) cross-linguistically, and attempt to provide instances of LDA that cannot involve any local relation, and must involve long-distance Agree. Finding such cases turns out to be harder than the literature suggests, as many instances of LDA analyzed in terms of Agree can be analyzed in some other way. But not all of them can. Some instances of LDA really offer strong support for Agree. Jochen Trommer: On Portmanteau Agreement Slides as .pdf file Portmanteau affixes are markers which in an intuitive sense ''replace'' two or more simple affixes. While many approaches to inflectional morphology assume the existence of portmanteau affixes without explicit justification (e.g. Lakämper & Wunderlich, 1998), Stump (2001) uses data from Swahili affix blocking to argue that they are empirically necessary. In this talk I focus on multi-argument agreement and defend a view which is implicit in the foundational work on Distributed Morphology (Halle & Marantz, 1993, 1994): Portmanteau affixes don't exist and apparent portmanteaus result from single overt markers accompanied by zero exponence. I show that even under restrictive theoretical assumptions there are no agreement affixes which must be treated as portmanteaus and many alleged portmanteau affixes which have to be analyzed as simple affixes to capture important patterns in their distribution. Finally, I address the recent claim by Nevins (2007) that portmanteaus are connected typologically to specific markedness configurations, and argue that these observations follow from the specific status of non-third person categories for secondary exponence. David Pesetsky: Undermerge and Russian case morphology Handout as .pdf file Fabian Heck & Marc Richards: A probe-goal approach to agreement restrictions and incorporation in Southern Tiwa Handout as .pdf file Southern Tiwa (ST, Tanoan) exhibits agreement with up to three arguments (ergative, absolutive, dative). This agreement is subject to certain restrictions. Moreover, there is a correlation between agreement restrictions and conditions on noun-incorporation in ST. The most explicit (and elegant) analysis we know is Rosen (1990), which makes use of a heterogeneous feature hierarchy and rules of association, familiar from autosegmental phonology. In our talk, we show how the full range of ST agreement and incorporation facts can be given a single, unified analysis within the Probe-Goal-Agree framework of Chomsky (2001). Gerhild Zybatow: Aspect in Russian and Verb Classes Karlos Arregi & Andrew Nevins: Ergative Proclisis in Basque: Wackernagel-driven Metathesis Slides as .pdf file Morphological metathesis, sometimes also called Local Dislocation, has been formalized in various ways. One important consequence of the formalism developed by Halle & Harris (2005) is that metathesis is generated by the same mechanism as partial reduplication and morpheme doubling. We demonstrate that a process of postsyntactic linear displacement of the ergative clitic in Basque to proclitic position should be modeled in terms of the Halle-Harris formalism, drawing evidence from the existence of doubling/circumclisis in certain varieties. The metathetic process is triggered by a word-internal 2nd position requirement on the auxiliary, otherwise filled by an absolutive proclitic. As doubling accomplishes this end as well as metathesis, both may persist as morphotactic repairs. Marc Richards: Defective Agree and Case Alternations This paper proposes a general schema underlying syntactic Case alternations based on the mechanism of defective probes (in the sense of Chomsky 2001), such that a defective head may value a different Case from its nondefective counterpart (cf. Rezac 2004). The resultant 'defective Case forms' are characterized by a range of well-known interpretive restrictions on argument encoding (definiteness-, animacy- and PCC-effects; scope freezing) - examples include Icelandic nominative objects, English expletive-associates, the Russian genitive of negation, the Hindi ergative, and the absolutive in Tanoan. These interpretive restrictions, and their relation to the EPP (optional vs. obligatory), are shown to follow from variable crosslinguistic association of the Person feature with, for probes, the Edge Feature of Chomsky 2005, and, for goals, different degrees of semantic prominence (as defined, for example, on a universal scale; cf. Aissen 2003). In this way, differences in form (Case-marking) have semantic consequences, with the various interpretive restrictions at the interface reducing to a single, common source: namely, formal violations of the Case Filter in the context of defective Agree. Back to Program Last Change: 02.06.2007
Harvard University Department of Linguistics University of Leipzig Institute of Linguistics University of Leipzig Institute of Slavic Studies Research Group Grammar and Processing of Verbal Arguments

Abstracts

Daniel Harbour: A Feature Calculus for Silverstein Hierarchies

Silverstein (1976) observed an important implicational relation between person and grammatical function. Consider the hierarchy below (a simplified version of Silverstein's):

1/2 > pronominal 3rd > human > other

According to Silverstein, if, in a language L, [+human] NPs (say) receive special marking when the direct object of the verb, then all hierarchically higher NPs (in this case third person pronouns and local arguments) receive special marking when they too are direct objects in L. Conversely, if pronominal third person arguments (say) receive special marking when external arguments in L, then all hierarchically lower NPs (in this case, [+human] and other arguments) receive special marking when they too are external arguments. I propose an account of these facts that extends Adger and Harbour's (2007) analysis of syncretisms frequently concomitant with Person Case Constraint in unrelated languages.

The core idea I will build on is that there are two different sources for phi-features on arguments in the syntax. On the one hand, heads minimally bear the phi-specification that their semantics demands. For instance, first and third person plural must both bear features for plurality, but only first person plurals must also bear a feature specification for person. On the other hand, heads may receive phi-features from the sister with which they are merged. Thus, third person arguments, for instance, differ in their phi-specification depending on whether they are merged with the verb or with a featurally contentful head such as v. Thus, an argument low on the Silverstein Hierarchy may, in virtue of the head that it merges with, receive a greater phi-specification than it intrinsically needs, and, conversely, an argument high on the hierarchy may endow the head with which it merges with more phi-features than it, the non-argument, intrinsically bears. On this conception, the special marking observed by Silverstein receives a straightforward analysis: it is the realization of the non-intrinsic phi-features that heads comes to bear in virtue of relations established by initial merge. The implicational relations (e.g., if [+human] NPs are morphologically marked as objects, so are all hierarchically higher NPs) follows from entailment relations between feature specifications.

Petr Biskup: Prefixes and T-Features

Abstract as .pdf file

Handout as .pdf file

Hakyung Jung: Genitive Subjects and Ergativity in North Russian

It has been argued that ergativity is syntactically reduced to a structure in which the case-marking of the external argument is ascribed to a functional node lower than the matrix tense, which permits the internal argument to check the Nom against the T or the Acc against v (Woolford 1997, Ura 2000, Anand & Nevins 2006 among others). In this respect the so-called impersonal passive in North Russian with a genitive subject and a nominative/accusative object resembles an ergative construction. This paper examines the underlying structure of the given construction and evaluates it adequately in light of a typological perspective. I suggest that ergativity is not confined to a particular case morphology but appears as a general case marking strategy tied to certain syntactic positions in many languages, where an oblique case(s) is utilized to mark a subject.

Gereon Müller: On Deriving Paradigm Economy Effects

It is shown that assuming instances of syncretism with argument encoding to be systematic in the unmarked case may significantly reduce the number of possible inflection classes that can be generated on the basis of a given inventory of markers, without recourse to specific constraints like Andrew Carstairs-McCarthy's Paradigm Economy Principle or No Blur Principle. If there is always one radically underspecified (i.e., elsewhere) marker per morphological domain, and if there is always one unique marker that is chosen in cases of marker competition, it turns out that there can be at most 2^n-1 inflection classes for n markers, independently of the number of instantiations of the grammatical category that the markers have to distribute over. The argument relies on the notion of marker deactivation combinations.

Cedric Boeckx: Isolating Agree

In this talk I examine various instances of long-distance agreement (LDA) cross-linguistically, and attempt to provide instances of LDA that cannot involve any local relation, and must involve long-distance Agree. Finding such cases turns out to be harder than the literature suggests, as many instances of LDA analyzed in terms of Agree can be analyzed in some other way. But not all of them can. Some instances of LDA really offer strong support for Agree.

Jochen Trommer: On Portmanteau Agreement

Slides as .pdf file

Portmanteau affixes are markers which in an intuitive sense ''replace'' two or more simple affixes. While many approaches to inflectional morphology assume the existence of portmanteau affixes without explicit justification (e.g. Lakämper & Wunderlich, 1998), Stump (2001) uses data from Swahili affix blocking to argue that they are empirically necessary. In this talk I focus on multi-argument agreement and defend a view which is implicit in the foundational work on Distributed Morphology (Halle & Marantz, 1993, 1994): Portmanteau affixes don't exist and apparent portmanteaus result from single overt markers accompanied by zero exponence. I show that even under restrictive theoretical assumptions there are no agreement affixes which must be treated as portmanteaus and many alleged portmanteau affixes which have to be analyzed as simple affixes to capture important patterns in their distribution. Finally, I address the recent claim by Nevins (2007) that portmanteaus are connected typologically to specific markedness configurations, and argue that these observations follow from the specific status of non-third person categories for secondary exponence.

David Pesetsky: Undermerge and Russian case morphology

Handout as .pdf file

Fabian Heck & Marc Richards: A probe-goal approach to agreement restrictions and incorporation in Southern Tiwa

Handout as .pdf file

Southern Tiwa (ST, Tanoan) exhibits agreement with up to three arguments (ergative, absolutive, dative). This agreement is subject to certain restrictions. Moreover, there is a correlation between agreement restrictions and conditions on noun-incorporation in ST. The most explicit (and elegant) analysis we know is Rosen (1990), which makes use of a heterogeneous feature hierarchy and rules of association, familiar from autosegmental phonology. In our talk, we show how the full range of ST agreement and incorporation facts can be given a single, unified analysis within the Probe-Goal-Agree framework of Chomsky (2001).

Gerhild Zybatow: Aspect in Russian and Verb Classes

Karlos Arregi & Andrew Nevins: Ergative Proclisis in Basque: Wackernagel-driven Metathesis

Slides as .pdf file

Morphological metathesis, sometimes also called Local Dislocation, has been formalized in various ways. One important consequence of the formalism developed by Halle & Harris (2005) is that metathesis is generated by the same mechanism as partial reduplication and morpheme doubling. We demonstrate that a process of postsyntactic linear displacement of the ergative clitic in Basque to proclitic position should be modeled in terms of the Halle-Harris formalism, drawing evidence from the existence of doubling/circumclisis in certain varieties. The metathetic process is triggered by a word-internal 2nd position requirement on the auxiliary, otherwise filled by an absolutive proclitic. As doubling accomplishes this end as well as metathesis, both may persist as morphotactic repairs.

Marc Richards: Defective Agree and Case Alternations

This paper proposes a general schema underlying syntactic Case alternations based on the mechanism of defective probes (in the sense of Chomsky 2001), such that a defective head may value a different Case from its nondefective counterpart (cf. Rezac 2004). The resultant 'defective Case forms' are characterized by a range of well-known interpretive restrictions on argument encoding (definiteness-, animacy- and PCC-effects; scope freezing) - examples include Icelandic nominative objects, English expletive-associates, the Russian genitive of negation, the Hindi ergative, and the absolutive in Tanoan. These interpretive restrictions, and their relation to the EPP (optional vs. obligatory), are shown to follow from variable crosslinguistic association of the Person feature with, for probes, the Edge Feature of Chomsky 2005, and, for goals, different degrees of semantic prominence (as defined, for example, on a universal scale; cf. Aissen 2003). In this way, differences in form (Case-marking) have semantic consequences, with the various interpretive restrictions at the interface reducing to a single, common source: namely, formal violations of the Case Filter in the context of defective Agree.

Back to Program

Last Change: 02.06.2007

Harvard University
Department of Linguistics

University of Leipzig
Institute of Linguistics

University of Leipzig
Institute of Slavic Studies

Research Group
Grammar and Processing
of Verbal Arguments