High Desert Linguistics Conference -V

University of New Mexico

November 1st & 2nd

2002 Presenter's abstracts

Catie Berkenfield, University of New Mexico

Putting Bodies into Embodiment: A Feminist Perspective on Metaphorical Structure

In this paper, I explore how current understanding of metaphor in cognitive linguistics, especially the metaphorical conceptualization of anger, is a product of historically situated methodologies in the field of linguistics, with links to the epistemology of rationalism. I suggest ways in which linguistic methodology in a feminist sociology of knowledge intervenes in and reconstructs our understanding of the role of the subject in metaphorical conceptualization. Specifically, I argue that a feminist perspective linking metaphorical structure, content, and use to local contexts of persons’ lived experiences yields a finer-grained analysis of metaphorical conceptualization than that of the cognitive account critiqued here. The ways in which this analysis shifts our understanding of metaphor include the theorization of multiple positions for viewing conceptualization, (while limiting the possibilities of viewing conceptualization in any non-principled way), through direct reference to material conditions of metaphorical use; and the theorization of conceptualizations that have previously escaped linguistic comment for falling outside of its "theoretical" purview, i.e., those with a sociological, hence, socially embodied foundation.

This feminist perspective systematically grounds conceptualization in the work occupation and gender of the subject—social characteristics that the cognitive account does not theorize. The methodology for researching metaphor from the feminist perspective allows for a better understanding of the links between the psychological (conceptual) and social (lived) components of human experience, and places these components in a socio-historical perspective of post-rationalism. I propose that a feminist methodology in cognitive linguistics research can be realized through an application of an embodiment principle--theoretical knowledge of metaphor will arise insofar as it is systematically structured by its relations to the material conditions of its production.

Close Window

Ysaura Bernal-Enríquez, University of New Mexico

Tesoro Perdido: Socio-historical Conditions in the loss of the Spanish Language of la Nueva México

Many studies across the Southwest of the United States show that Chicano Spanish is not being transmitted to the youngest generations. In a project with 201 native speakers of New Mexico and Southern Colorado Spanish, ranging in ages from 18 to 96, proficiency, sociolinguistic, socio-demographic, and socio-historical data are analyzed in order to understand the relationship between attained proficiency and patterns of language use in this non-immigrant variety of Southwest Spanish. More specifically, the project explores the early childhood conditions for bilingual development among three different age groups: how the introduction of a societally-dominant second language at different stages of first language development affects both current use and the ultimate proficiency in the first and second languages. This paper will summarize the findings of the project and propose its implications for the roles of the home, the school, and the community in revitalization so that Chicano Spanish is once again successfully transmitted to the next generation. Language of presentation will be English, with quotes from consultants in Spanish.

Close Window

Esther L. Brown, University of New Mexico

New Mexican Spanish /s/: ¿Qué pa(h)a?

Syllable final /s/ is perhaps the most studied phonological variable in Hispanic linguistics. In sharp contrast to the plethora of studies on /s/ in final position, there is a dearth of empirical studies on initial /s/. In this study, I compare frequencies and constraint hierarchies for /s/ aspiration and deletion word initially (señor), word medially in syllable initial position (pasa), word medially in syllable final position (este), and word finally (vamos), in data from New Mexican Spanish. Contrary to the prevailing assumption that /s/ reduction diffuses from syllable final to syllable initial position (Lipski 1999), I find a higher frequency of reduction word medially in syllable initial position (pasa) than in syllable final position (este). Comparisons of these reduction rates, as well as constraint hierarchies determined from variable rule analyses, fail to provide evidence of a unified process of reduction spreading from word final to medial and initial position. Additionally, the present study looks beyond syllable and word boundaries to include frequent lexical items and combinations with high transitional probabilities known as "chunks" (Bybee 2000) in order to explore the effects of phonotactic patterns in /s/ reduction. Certain phonotactic patterns (such as /asa/ or /ose/) that occur word initially, medially, and finally deviate significantly from average reduction rates for each respective word position, illustrating a promising avenue of inquiry to examine the intersection between lexical and phonological effects in sound change.

Close Window

Matthew Burdelski, University of California, Los Angeles

Reported speech in Japanese parent-child interaction

When human beings interact they often find occasion to embed their talk with the talk of others in the form of reported speech. Reported speech can be defined as "a message belonging to someone else" (Volosinov, 1973: 201) or "speech which purportedly re-presents another specific speech event" (Lucy, 1993: 2). Within his analysis of participation frameworks, Goffman (1981: 150-151) pointed out that parents socialize children early on in this "self-dissociated" talk when, for example, during play they "speak for" objects in the child's surround such as stuffed animals or dolls.

Taking Goffman's observations as a point of departure, I searched my own data of Japanese parent-child interaction for examples of reported speech. The data were collected in and outside the homes of four families in Japan with one son under the age of four years (actual ages 1;9-4;0). In order to guide the present inquiry, I asked the following questions: Do parents report the speech of others when speaking to their children? If so, then in what contexts, and for what communicative purposes do parents use reported speech? In addition, following Wootten's (1997) analysis of children's participation in and understanding of conversational sequences, I asked: How do young children respond verbally and nonverbally following a parent's use of reported speech?

The analysis is qualitative and analyzes conversational sequences that have been transcribed from both video and audio taped data. For example, in the following excerpt (see APPENDIX), father, son (Ken: 2;2), and a neighborhood girl (Ami), are in a playground. Ken is currently holding a ball that belongs to the girl, and father had requested that he give it back to the girl. In this excerpt, when the girl announces that she is going home (line 5), father quotes the girl's speech (line 7), which escalates the urgency for Ken to return the ball. Ken responds by running away (line 8) and by articulating his defiance (line 10), which seems to display an understanding of the escalated urgency that is indexed through the reported speech.

This presentation will present several examples of this type of reported speech, and suggest that reported speech, while pulling what was said by a third party into the child's present consciousness, serves to entice, invite, and otherwise direct a child towards some particular course of action. This type of directive can be considered an "indirect" directive, as analyzed by Clancy (1986) in her investigation of acquisition of communicative style in Japanese. This type of directive complements other directive strategies. The present analysis also builds on this previous work by focusing on children's verbal and nonverbal responses to utterances embedded with reported speech.

Close Window

Li-Hsiang Chang, University of New Mexico

Subjective Structures of the Mandarin Morpheme LE in Conversational Discourse

This is a pioneer study of the Mandarin morpheme LE based on spontaneous spoken conversational data. Previous researchers have identified the morpheme LE in Mandarin Chinese as a perfect (S-le) or perfective marker (V-le). In addition to the morpheme LE having the meaning of perfect or perfective, this study also manifests a brand new ‘pattern’ of the use of LE, which is identified as C-le (i.e. comment LE). This study is primarily concerned with the subjective use of S-le and C-le and their subjective structures in conversational discourse. According frequency analysis of collocation patterns of S-le and C-le utterances, this paper indicates that high frequent items of LE utterances are more subjective than the low frequent one; that is, the preferred subjective constructions of LE utterances are the patterns that appear most frequently in discourse. This study also suggests that subjective construction of S-le or C-le is used to signal a speaker epistemic or evaluative stance in discourse. The stance use of LE construction is highly associated with social interaction and/or action the speaker is engaged in. In other words, the main function of LE in interactional conversation is to emphasize the speakers’ viewpoints through a cooperative effort between speaker and hearer (i.e. interaction). The fact that the data from this study is taken from interactional conversation suggests that subjective usage must be taken into account to the grammar of the Mandarin morpheme LE. This is in favor of a functional analysis of language and emphasizes the importance of understanding the relationship of frequently recurring patterns in discourse and the emergence of grammar as well as the significant effects of the role of speaker point of view on grammar.

Close Window


Masanori Deguchi, Indiana University

The Source of Quantifier Scope

 [1] Introduction: The so-called "quantifier scope" is infamous for the subtlety of grammaticality judgment and the involvement of various non-syntactic factors. In discussing Japanese sentences involving quantificational NPs (henceforth QNPs), I will first investigate such non-syntactic determinants and then demonstrate that the dichotomy between background and foreground plays a crucial role in computing the interpretations of sentences with QNPs. In this paper, I argue that a QNP belonging to the background acts as an "anchor" and gives us "the impression" of having scope over other QNPs. To further support this stance, I will discuss some problems that any theory adopting quantifier raising (QR) cannot overcome, at least without paying rather significant conceptual costs.

[2] Fluctuating judgments: Quantifier scope seems sensitive to focusing effects. The contrast between (1a) and (1b) for example shows that the inverse scope reading becomes much easier when the subject QNP is narrow-focused (indicated by CAPITALS). As is explicated below, it is quite difficult if possible for an approach based on QR to capture such a wavering judgment without stipulations.

[3] Proposal: I will argue and show that a QNP gives "the impression of taking high scope" when it is part of the background (cf. Bolinger (1978) observes that a fronted wh-phrase in the English multiple-wh construction belongs to the topic of the topic/comment division). Suppose that the subject is the default position for background information (e.g., topic). The proposed analysis then predicts the subject QNP to scope over the object QNP by default. On the other hand, if the subject QNP is focused and thus foregrounded, the object QNP is expected to take scope over the subject QNP. These predictions indeed precisely coincide with the interpretations in (1a) and (1b), respectively. The same point can be also demonstrated with a use of the focusing particle ~mo 'as many as ~' as shown in (2). As predicted, two interpretations are available with (2a) depending on which QNP is construed as background. On the other hand, if the object QNP is foregrounded with the particle ~mo as in (2b), only the subject QNP can be interpreted as background and the inverse scope becomes unavailable. This is also a correct prediction. Given these, I therefore argue that a QNP in background functions as the anchor of the information structure and is "felt" to be scooping over other QNPs. On the other hand, QR-based analyses would be forced to make an ad hoc exception if the empirical facts discussed above were to be accommodated, e.g., exempting QNPs from undergoing QR when and only when they are part of the foreground (e.g., focused QNPs and QNPs with ~mo). In light of this, QR-based analyses need serious reconsideration because making QR available only to background QNPs would be speculative at best.

[4] Further support: There is another piece of evidence that favors the proposed analysis over QR-based approaches. There appear to be two types of QNPs that are lexically specified for the background/foreground value. The QNP dareKA 'someone' in (3a) is lexically foregrounded and thus never scopes over the object QNP. On the other hand, the expression darekasan 'someone' as in (3b) is lexically specified for background (in a similar sense that a which-phrase is inherently D-linked (Pesetsky 1987)) and always takes scope over other QNPs regardless of the logical environments (cf. Type I quantifiers of Hornstein (1984)). This fact is problematic for QR-based analyses because QR would have to treat these two types of QNPs differently; e.g., QR only applies to one type, not the other.

[5] Conclusion: We first noted that the "fragile" nature of quantifier scope cannot be captured without referring to the information structure of sentence. After pointing out problems associated with QR, I claimed that what is often perceived as "high scope" is after all the effect of backgrounding, rather than an effect of QR. Some of the issues that follow from the proposed approach include (i) why the surface scope reading is almost always available; (ii) why even prosody affects interpretation and it does so systematically; (iii) why pragmatic/discourse information affects the interpretation of sentences with QNPs.

Close Window


Paul Edmunds, University of New Mexico

Shakespeare, Dickens, and Christie: A Diachronic Study of English Contractions

 This paper reflects language change through the study of English contractions through the work of British texts, specifically of the fictional genre. This study examines diachronic change of English contractions for the verbs be, have, will, and would with personal pronouns from the time of Shakespeare (early 17th century), Charles Dickens (early 19th century) and Agatha Christie (20th century). Analysis of the corpora of six Shakespeare plays, one Dickens text, and two Christie texts shows that relative frequency of use of contractions increased diachronically in the majority of categories analyzed. Results show that the verbs be and have tended to show more contraction with third-person pronouns, while the inherently more modal will and would showed more contraction with first-person pronouns. This increased frequency of contraction is likely due to the influence of spoken communication on the works of current literary authors. British authors were chosen due to the fact that they are some of the first writers to have used contractions in written texts, according to the Oxford English Dictionary. Large corpora for each author help account for differences due to the variety of writing styles.

Close Window

David Eddington, University of New Mexico

Issues in modeling language processing analogically

Exemplar-based models of language assert that linguistic processing involves analogy to past linguistic experiences stored in the mental lexicon. This study explores how three factors influence the predictions made by exemplar-based simulations of linguistic processing. Three questions are posed as they relate to such simulations: 1) Is type frequency or token frequency a better predictor of outcomes? 2) What is the optimal way of aligning the variables in the database so that the most relevant analogs are found? 3) Are there significant differences between representing variables as phonemes versus representing them in terms of distinctive features? Spanish stress assignment and English past tense formation served as the linguistic phenomena on which these issues were tested. The results suggest that type frequency is a better predictor of outcomes, although simulations using token frequency were most successful when only middle frequency words were included. Several methods for aligning variables in the analogical database are discussed. The dual-alignment method has advantages for the English past tense task, but not in predicting Spanish stress. In the Spanish task, strict phonemic representation of words demonstrated no advantage over feature representation. However, phonemic representation produced better results than distinctive features in predicting the English past tense.Key words: Analogical Modeling of Language, distinctive features versus phonemic representation, English past tense, Spanish stress assignment, Tilburg Memory-based Learner, token frequency, type frequency, variable alignment.

Close Window


Patrick Farrell, University of California, Davis

The Conceptual Semantics of Cognitive Verbs in Brazilian Portuguese

 The verbs cansar ‘tire’, ver ‘see’, and gostar ‘like’ represent three classes of cognitive verbs in Brazilian Portuguese (BP), which are distinguished by the syntactic properties of their complements. Employing ideas from cognitive grammar (e.g., Langacker, 1987, 1991) and a historical semantic analysis, this paper shows that the varying syntactic properties of these verb classes are motivated by conceptual semantic differences. All three verb types have their experiencer as subject. The stimulus argument of ver is a bare noun phrase with all typical direct object properties (Farrell 1989), including the ability to be expressed as a clitic pronoun (1b), to be construed with an adjectival secondary predicate (1c), and to undergo passivization (1d), tough movement (1e), and bare NP topicalization(1f).The stimulus of cansar is marked with the preposition de ‘from, of’ and is a typical oblique, with no direct object properties, as shown by (2b-f). Although the stimulus argument of gostar is marked with the preposition de, unlike the stimulus of cansar it behaves like a direct object with respect to topicalization, tough movement, and adjectival secondary predication—but not with respect to cliticization and passivization, as shown by (3b-f). As noted by Lipski (1975), the verb meaning ‘like’ in Spanish and Portuguese developed in different ways from the Latin verb meaning ‘taste’. The most common Spanish reflex is a construction in which the experiencer is dative—but with certain subject properites (e.g., González 1985, Cuervo 1999)—and the stimulus is a nominative subject (A Paulo le gustan flores ). This outcome as well as the BP alternative are attributable to the conceptual semantics of the original verb. The concept ‘taste’ involves a force-dynamic interaction between two participants. The taster acting on the tastee, by means of the tongue, is the primary action. There is a conceptually subliminal secondary action, however, wherein the intrinsic properties of the tastee affect the taster, giving rise to a gustatory sensation . By routine metaphorical extension, the verb gustar/gostar came to designate the event associated with the prototypical sensation resulting from tasting, i.e., pleasure. Crucially, the force-dynamic interaction between experiencer and stimulus is preserved in the derived concept. Spanish opts for the ‘liker acts on likee’ action as the basis for the nominative subject; BP, on the other hand opts for ‘liker acts on likee’. The preposition de marks the stimulus as the source of pleasure, which consequentially abstractly acts on the experiencer, as in the case of cansar. Unlike with cansar, however, the experiencer also acts abstractly on the stimulus, for which reason the stimulus has some of the properties of a typical direct object.

Close Window

Samira Farwaneh, University of Arizona

Housewives, Harems, and Houris: Asymmetries in Gender Paradigms

Lakoff’s (1975) impressionistic yet influential work generated an upsurge in research on language and gender issues, primary among which is the detection of sexist language usage in English and other languages. Evident throughout this literature is the assertion that sexist language usage, i.e., the overt or covert expression of gender bias, manifest itself via a variety of linguistic behavior including vocabulary selection, naming practices, use of euphemisms, selection of title terms, semantic derogation, generic use of the masculine form, among other phenomena.

In this paper, a careful evaluation of gender differentiation within the nominal system in Arabic is undertaken focusing on asymmetries in cross-classification, wherein one gender classification is attested without its counterpart, or where the two forms exhibit inequitable semantic connotations. Such gaps are exemplified in English by forms like ‘spinster’, which lacks a masculine equivalent, or ‘master’ whose feminine counterpart ‘mistress’ has been denigrated.

In a gender-marked language like Arabic, wherein gender differentiation is displayed through lexical dichotomies, e.g., [rajul/imra?a] ‘man/woman’, [bint/walad] ‘girl/boy’,

or through a productive affixation process, e.g., [mudarris/mudarris-a] ‘teacher’ one would expect all nominal and adjectival forms to fall neatly in pairs creating a perfectly symmetrical paradigm, except where physiology dictates otherwise, e.g., [Haamil] ‘pregnant’. However, I have found several gaps that are inexplicable in terms of biological differences. This paper examines such asymmetries through consideration and evaluation of three types of linguistic behavior: naming practices, terms of address preferences, and coinage of status (religious or social) terms.

The data collected reveal three types of asymmetries: Cases where both gender lexemes are theoretically possible yet only one is attested, e.g., [muHallil] *[muHallil-a]

‘enabler-Masc.’ and [Huur] ‘houris/celestial maidens’; cases where both gender forms exist with semantic derogation of the feminine, e.g., [rab bayt] ‘lord/master of thee house’ vs [rabb-at bayt] ‘housewife’ and [maHram] ‘instrument of prohibition-Masc.’

vs. [Hurm-a/Hariim] ‘object of prohibition-Fem.’; and cases where the lexeme is morphologically gender-neutral but syntactically masculine yet used exclusively as a feminine referent, e.g., [Hanaan] ‘tenderness-Masc.’ used exclusively as a girl’s name.

By analyzing the correlation between the morphological structure and semantic features of a gender term and the referential power it acquired in the culture of its speakers,

I demonstrate how asymmetries in cross classification reveal the underlying gender bias and the gender role stereotypes pervasive in Arab societies. Interestingly, a search of the Quran and Hadith (the main sources of Islamic Jurisprudence) databases for the gender differentiated terms under investigation revealed no results, which leads to the hypothesis that social forces more powerful than and probably foreign to religion must have played a crucial role in the construction of gender in the Modern Arab World.

Close Window

Franca Ferrari, New York University

The Perfect Correspondence

In the Romance literature, it has been often claimed a lack of a one-to-one correspondence between gender features and the gender markers. For instance, in the case of Italian, it has been considered non-true the claim that all nouns ending with the gender marker [-o] are masculine, e.g. "baci-o" (kiss), and those ending with the gender marker [-a] are feminine, e.g., "cas-a" (house). This claim is based on the presence of a group of exceptions, i.e., feminine nouns such as "mano" (hand) and masculine nouns such as "poema" (poem). These exceptions led to the introduction of the notion of word markers.

According to Harris (1991a, 1991b), word markers are "the phonological signature of arbitrary classes of words". They are present on all lexical items with the exception of finite verbs. Word markers are used to form declension classes and not exclusively noun classes. In Italian, for instance, on the base of the number of word markers, one can devises four declensional classes (nouns ending with word markers [-a], [-o], [-e] plus a catch-all class for residue nouns) in a system with two grammatical genders (masculine and feminine). The presence of word markers implies that the relation between gender features and gender markers is arbitrary.

In this talk I will argue against the very existence of word markers in Italian. An accurate analysis of Italian nouns reveals that the so-called-exceptions cannot be used as an argument in favor of the existence of word markers. The elimination of word markers re-appropriates the gender markers of a morphological content. Gender markers, in fact, become the overt expression of the lexical gender feature. Moreover, it predicts that the number of noun classes is proportional to the number of gender features. This means that in Italian, firstly, there are two noun classes, the feminine and the masculine class, and, secondly, that the claim that all nouns ending in [-a] are feminine, whereas all nouns ending in [-o] are masculine is correct. I will show that these facts about Italian gender are correct by demonstrating the following.

a. I will demonstrate that the end vocalic segment of the so-called exceptions does not behave as a word marker. For instance, contrary to word markers, which decline in plural

([-o] sg. > [-i] pl. and [-a] sg. > [-e] pl.), the end vowel –o of the masculine exceptions is invariable as it does not change in the plural (see data 1). Whereas the end vowel –a of the feminine exceptions declines using the masculine plural word marker [-i] rather than the expected feminine [-e] (see data 2).

Moreover, the insertion of the diminutive affix " -in- " between the nominal stem and the word marker does not change the nature of the word marker (see data 3). Contrarily, in presence of a diminutive the feminine and the masculine exceptions change their end vocalic segment into the gender marker corresponding to their inherent gender feature (see data 4).

b.Finally, I will demonstrate that the epenthetic nature of end vowel [-e] allows the insertion of nouns ending in [-e], which can be either feminine or masculine, in one of the two existing noun classes.


1. la moto > *le moti, * le mote, *i moti
2. il poema > * i poeme
3a. cas-a > diminutive form > cas-in-a (little house)
3b. baci-o > diminutive form > bac-in-o (little kiss)
4a. mano > diminutive form > man-in-a (little hand)
4b. poema > diminutive form > poem-in-o (little poem)

Close Window


Kareen Gervas, University of Southern California

On the interaction of Syntax and Discourse in Spanish Relativization

Previous studies have explained the process of relative clause formation from a typological approach (Keenan and Comrie (1977) and Comrie and Keenan (1979)). These authors established an Accessibility Hierarchy based on cognitive difficulties involved in the relativization of various grammatical roles.

From a discourse level perspective, Fox and Thompson (1990) showed that a number of factors present in the interaction determine grammatical choices that the speakers make when using English relative clauses.

Following Fox and Thompson’s variationist approach, this paper discusses patterns of relativization of nominal phrases (NPs) in the roles of Subject (S) and Direct Object (DO) in conversational Mexican Spanish (Lope Blanch 1971, 1976). The purpose is to explain semantic/pragmatic factors that underlie Spanish relativization.

Four types of structures are analyzed, based on the grammatical role of the NP antecedent of the relative clause (underlined) and of the coreferent NP relativized (within brackets). Subject relatives are NPs relativized in the role of Subject, whereas Object relatives are NPs relativized in the role of Direct Object. I illustrate these types in 1-4:

(1) S [S (Subject relative)]

'...el padre [que me dirigía en la clase] dijo:...’ (Lope Blanch 1971: 130).

'...the priest [that guided me in the class] said:...’

(2) S [DO (Object Relative)]

'..., esas compañías [que te acabo de decir]...tienen la calidad;...’ (Lope Blanch 1971: 39). '...those companies [I just mentioned to you]...they have quality;...’

(3) DO [S (Subject Relative)]]

'Entonces, yo ya empecé a meter un albañil, un peón, [que me empezara a desmontar...a quitar la piedra]’ (Lope Blanch 1976: 191).

'Then, I started bringing a bricklayer, a worker, [that would start removing the stone for me]’.

(4) DO [DO (Object Relative)] 'Y una vez fui yo para recoger unas cosas [que yu había dejado allá...] (Lope Blanch 1976: 72). 'And once I went to pick up some...some things [that I had left there...]’.

I claim that the variation shown in (1)-(4) is influenced by factors such as the humanness of the NP antecedents, the speakers’ strategies to make NP antecedents relevant in discourse (grounding) and the functions of relative clauses with respect to their antecedents. The statistical results show a correlation between types of grounding and the use of Subject relatives and Object relatives. The data also reveal that Subject relatives and Object relatives exhibit different patterns in terms of the functional roles that they achieve in the interaction.

Close Window

Gabriela del Carmen Gonzalez Gonzalez, Universidad de Colima

Modalidad Epistemica en el Verbo Andar

El verbo andar ha experimentado un proceso de gramaticalización que lo ha integrado al conjunto de los verbos auxiliares del español; además, de su función auxiliar, se presenta en ciertas expresiones que permiten considerarlo como verbo modal. En este documento, intento explicarlo como modal epistémico desde la perspectiva de Langacker (1991a).

Andar como verbo modal incluye los contenidos relativos a actitudes que afectan a la perífrasis completa. Dichas actitudes surgen gracias al involucramiento del hablante con lo que dice y se gramaticalizan gracias a la trayectoria difusa que está en la base de la conceptualización del verbo.

Cuando un hablante emplea el andar modal, sólo tiene ciertas pistas acerca de lo sucedido, no es testigo de los hechos. Tal información la obtiene por rastreo conceptual. El hablante hipotetiza acerca de la realidad; por consecuencia, las expresiones con andar modal no son parte de la realidad inmediata, sino que se ubican en la realidad potencial o en la realidad proyectada, como se puede apreciar en los siguientes ejemplos:

Realidad potencial: a) Nicolás anda en líos de faldas.

Realidad proyectada: b) Laura se anda casando.

En a) quien habla tiene cierta información sobre lo acontecido, no necesariamente precisa; surge de datos previos, de un proceso deductivo o de elucubración personal. Lo que dice el hablante no forma parte de su realidad inmediata, hay distancia entre el evento y él; ello genera una expresión cuyos contenidos son hipotéticos.

En b) la acción se proyecta hacia el futuro y expresa simultáneamente una intención y el desarrollo de un proceso. Cuando se emplea esta expresión el hablante puede implicar sólo la intención, o incluir también el proceso y los trámites correspondientes. La idea de proyección al futuro permanece subyacente.

Así, se puede apreciar que el verbo andar es un modal epistémico, que se refiere más a lo que las cosas parecen que a lo que son. El hablante se expresa sin tener evidencia de lo que dice, sólo externa sus creencias y suposiciones, a partir de su conocimiento de la realidad, el cual es extendido hacia otras situaciones aún no constatadas. En la mayoría de los casos, las expresiones incluyen las actitudes del hablante hacia sus contenidos.

Close Window

Lisa J. Graham, Washington College

Renovation in Clause Level Relations in Old High German

Grammaticalization as a term refers to a framework within which to account for language phenomena, to that part of the study of language that focuses on how grammatical forms and constructions arise. From a historical perspective, grammaticalization investigates the sources of grammatical forms and the typical pathways of change that affect them. In the development of subordinators, grammaticalization must be considered as a broad parameter pertaining to the reanalysis of discourse patterns as grammatical patterns and of discourse-level functions as sentence-level, semantic functions.

Old High German is the collective name for a number of dialects spoken between c. 770 - c. 1050. Otfrid von Weissenburg's 9th century Evangelienbuch is an Old High German verse rendering of the gospel story that holds a place of prime importance in the meager store of extant German literature from the early Middle Ages. Writing in the vernacular, Otfrid’s aim was to clarify, interpret and instruct, thereby satisfying the needs of those who were precluded from the study of the Vulgate by the difficulties in Latin. Otfrid interposes explanatory phrases, treating the Bible in its own context and applying its wisdom to wider issues. He interrupts the flow of narrative with little asides, helpful explanations and moral ruminations. Otfrid's lengthy work is particularly suited to an examination of the development of subordinators as he interposes explanatory phrases, adds commentary and interpretation, thus establishing a need for subordinating causal conjunctions is creating his discourse.

While dependent clauses exist in the oldest Indo-European languages, subordination is not highly developed in the oldest Germanic dialects. Investigation of conjunctions from a diachronic point of view reveals the existence of a relatively fixed system of conjunctions. Although the system is already found in OHG, the individual conjunctions making up this system undergo numerous changes during their history. One of the most important conjunctions in early OHG developed from the demonstrative nominative and accusative neuter pronoun, thaz ‘that’. The pronoun originally belonged to the main clause but referred to the contents of the following clause that had the characteristic of a dependent assertion. This paper will explore the development of this causal conjunction within the discourse of Otfrid's Evangelienbuch as the deictic component of this demonstrative is weakened in the first stage of grammaticalization while expressivity and scope of the developing conjunction increases.

Reanalysis of discourse patterns as grammatical patterns and of discourse-level functions as sentence-level, semantic functions reveals the development of the various syntactic functions of the thaz-clause, from complement to causal clause. As the thaz-construction reigns over almost the entire spectrum of semantic variation of subordinating clauses with a structural and logical relationship between the subordinate and the main clause in Otfrid's work, it will be shown that the pronominal form expands to include a complementizing function that subsequently undergoes a number of grammaticalization phenomena and pragmatic strengthening of informativeness that leads it outside of its original domain. Also evidenced in this processes of generalization, categorical downshifting and renovation is an increased strengthening of grammatical meaning marking grammatical relations in the development of conjunction from deictic demonstrative.

Close Window


Mie Hiramoto-Sanders, University of Hawaii at Manoa

Variations in the Japanese Lexical Strata

This paper studies the Japanese lexical strata based on vocabulary of different categories. Some scholars (i.e., Shibatani 1990, Ito and Mester 1995) separate Japanese vocabulary in three basic categories, namely the native or Yamato-kotoba (Yamato, hereafter), Chinese loanwords from the 5th century (Sino-Japanese, hereafter), and other loanwords from western languages (Foreign, hereafter). Certain phonological rules show variation of their applications depending on the lexical categories of the Japanese vocabulary. These phenomena are explained well within lexical strata and an idea of lexical phonology. For example, voicing assimilation in compounds (Rendaku Rule) primarily takes place in Yamato stratum but not in Sino-Japanese nor in Foreign strata. Therefore, the rule application seems to be stipulated at the earliest stratum of the lexical category.

Japanese has borrowed from other languages besides Chinese and western languages in its history. Some of the early loanwords have nativised and behave like Yamato vocabulary in the rule application. For example, an Ainu loanword shake ‘salmon’ and a Sanskrit loanword kawara ‘roof tile’ both go through Rendaku Rule and gain [+voice] feature in compound forms as in shio-jake ‘salted salmon’ and oni-gawara ‘a decorative roof tile in a shape of devil’s face’. The rule application exceptionally takes place in some Sino-Japanese and Foreign vocabulary, as in oo + chawan ‘big + rice bowl’ àoojawan ‘big bowl’ and iroha + karuta ‘Japanese alphabets + playing cards’ à iroha-garuta ‘playing cards of Japanese alphabets’.

In many cases, the Japanese native speakers are not aware that words such as shake, kawara, chawan or karuta are words of the foreign origin. Based on the observation and my own native speaker intuition, it seems as the Japanese native speakers’ psychological reality may be appealing to the Japanese lexical strata and variations in the rule application.

Close Window

Harry Howard, Tulane University

Spanish mood as contrast enhancement

Following Bolinger (1968) and Terrell & Hooper (1974), Lunn (1995) characterizes indicative mood as marking a clause as an assertion and subjunctive mood as marking a clause as a non-assertion. Lunn points out three reasons why a speaker may view a clause as unworthy of assertion: (1) because the speaker has doubts about its veracity, e.g. Dudo que sea buena idea; (2) because the proposition is unrealized, e.g. Necesito que me devuelvas ese libro; or (3) because the proposition is presupposed, e.g. Me alegra que sepas la verdad. Lunn builds on Lunn (1989) to further refine this characterization by drawing on Relevance Theory, see Sperber & Wilson (1986), to claim that potentially assertable information must have two qualities, “… in the speaker’s opinion, it must be both reliable as to truth value and informative as to news value.” (Lunn 1995:430) Thus (1-2) fail the first desideratum, and (3) fails the second. In this way, Lunn can unite the seemingly aberrant (3) with the prototypical usages of (1-2).

In this paper, Lunn’s description is cast in a neurologically-inspired framework which attempts to explain why language should be sensitive to assertability in the first place, and how the brain creates this sensitivity. In particular, we claim that the indicative/subjunctive dichotomy should be assimilated to the notion of contrast enhancement among alternatives: the indicative marks maximal contrast, while the subjunctive marks minimal contrast.
Contrast enhancement has long been observed in sensory systems and is posited to underlie much cognitive functioning as well. Neurophysiologically, it is instantiated by the mechanism of lateral inhibition. In a nutshell, lateral inhibition is a negative connection among neural units that encode similar input that ensures that only one of them will respond to a given stimulus; it is counterpoised to excitation, a positive connection among units that encode dissimilar input that ensures that many of them will respond to a given stimulus.

Consider by way of illustration the network in Fig. 1. It contains three groups of similar units, 3 nouns, 2 verbs, and 3 prepositional phrases. Within the groups, connections among units terminate in balls, signifying inhibition; between the groups, connections among units terminate in points, signifying excitation. Such a network can be given mathematical form through an algorithm known as Interactive Activation and Competition (IAC, McClelland & Rumelhart 1981) and used to simulate what happens when some sub-network is activated. For instance, turning on the three units that encode “María va al trabajo” and allowing their activation to circulate through the network for 13 iterations of the IAC algorithm results in Fig. 2 and 3, which are distinguished by the setting for lateral inhibition: high in Fig. 2 and low in Fig. 3. In Fig. 2, the high setting allows the target clause to be clearly separated from its background alternatives – the contrast between them has been enhanced by the network dynamics. We claim that this is the neurological function of the indicative. In Fig. 3, on the other hand, the low setting does not allow the target to be separated from its background alternatives – in fact, there are points at which some of the background attains a higher activation than the target. The contrast between the two winds up being obscured by the network dynamics. We claim that this is the neurological function of the subjunctive.
Thus the simple IAC network in Fig. 1 distinguishes between assertion and (1) doubt and (2) non-realization mentioned by Lunn above. It can be modified to account for (3) presupposition, by assuming that the presupposed (i.e. background) material is equivalent to the target material, thereby reducing the contrast between the two to zero. Finally, this mechanism automatically evaluates the target with respect to some background material, as postulated in the model-theoretic analysis of Villalta (2000), while retaining considerable neurological realism.

 Figure 1. A network with local inhibitory and long-distance excitatory connections

Figure 2. Maximal separation of target from alternatives -> assertion/indicative

Figure 3. Minimal separation of target from alternatives -> non-assertion/subjunctive


Close Window


Lingyun Ji, University of Southern California

Language Variation between the Two Registers: Government Press Conferences and Public Conversations

Government press conferences are a special register, which is similar to public conversations in that both the former and the latter happen before the public. The two registers, however, are much different from each other in language style. This paper studies the stylistic variations between the register of government press conferences and that of public conversations by employing the model of stylistic variation developed by Biber & Finegan (1989a, b). The paper focuses on the first empirically defined dimension of linguistic variation (Biber & Finegan, 1989a, b): "informational vs involved production".

The data in the paper are collected on line and consist of four Chinese texts: transcriptions of two Chinese government press conferences and transcriptions of two public conversations.

The paper analyzes the stylistic variation of the two registers by comparing the occurrence frequencies of seven features in the dimension "informational vs involved production" in Biber and Finegan’s (1989a, b) model of factor analysis. The features that are investigated in the paper include: nouns, private verbs, 2nd-person pronouns, demonstrative pronouns, 1st-person pronouns, general emphatics, and general hedges.

Different distributions of the features in the four Chinese texts clearly display the stylistic variations between government press conferences and public conversations. The paper shows that government press conferences are more informational than public conversations. Government press conferences use much more nouns, especially much more complicated nouns, than public conversations. Also, sentences in government press conferences are noticeably longer than those in public conversations. Furthermore, public conversations tend to choose more private verbs, pronouns, general emphatics and general hedges, which occur less frequently in government press conferences. Particularly, general hedges occur only once in government press conference texts in our data.

The paper also discusses the reasons why the two registers differ much from each other in style. Stylistic variations displayed in the data can be accounted for if we consider the differences of the social and linguistic contexts in which the two registers are used.

The representation forms of the features that occur in the four Chinese texts are given in appendix, so that future research on register variation in Chinese texts can follow the current research.

Close Window

Joel Judd, Adams State College

The Language of Perceptual Control

Since the appearance of Behavior: The Control of Perception (Powers, 1973), the application of cybernetic control theory to the realm of learning and behavior has provided intriguing insights and alternatives to prevailing perspectives in such fields of study as sociology, psychology, child development and linguistics.

One of the more promising implications for linguistics is the unified picture offered by that Perceptual Control Theory’s (PCT) hierarchical levels of perception. Taken as a whole this hierarchy could account for the observable range of complexity in human language. A PCT hierarchy also describes a more satisfactory picture of the relationship between language and perception in general.

This presentation will explain the basis for viewing behavior as the control of perception, not as the outcome of linear cognitive processes or stimulus-response mechanisms. A theory, model, and mechanism for perceptual control will be offered. Then, language will be mapped onto an 11 level hierarchy of perception in order to explain how linguistic behavior can be seen as just another form of perceptual control, albeit in humans a highly specialized one. Some avenues for continued investigation will be outlined.

Close Window


Masumi Kai, Visiting Scholar at Harvard University

The double 'wa' Sentence in Japanese

The aim of this paper is to examine the meaning and the function of the second ‘NP wa’ in the double ‘wa’ sentence: the ‘X wa Y wa P’ construction.

It is claimed that ‘NP wa’ has two meanings as ‘topic’ and ‘contrastive’ (Mikami 1963, Kuno 1976). Based on this distinction, it is suggested that ‘NP wa’ near the predicate or ‘NP wa’ which is originally object or oblique case tend to have the contrastive meaning (Aoki 1992, Noda 1996, Ichikawa 1988). However, in this paper, we shall show examples which do not have the contrastive meaning on the second ‘NP wa’. Moreover, the interesting phenomena is, some sentences have a touch of contrastive meaning on the first ‘NP wa’ rather than the second ‘NP wa’.

Logically there are four combinations about the meaning of ‘X wa’ and ‘Y wa’ in the ‘X wa Y wa P’ construction.

Type1. Contrastive – Contrastive

Type2. Contrastive – Topic

Type3. Topic – Topic

Type4. Topic- Contrastive

However, Type1 does not exist because of the restriction that one sentence can only have one contrastive meaning; therefore we examine Type2 to 4.

The ‘X wa P wa P’ construction becomes Type2 or 3, if ‘Y wa’ is introduced as follows:

1. Y is mentioned in the previous sentence.

2. It is sequentially mentioned in the following sentence with particle ‘wa’ (discourse topic).

And here, ‘X wa’ is a subject (syntactic topic) or an upper concept of Y.

In short, ‘Y wa’ in the ‘X wa Y wa P’ construction can be a topic, if it is the "discourse topic" in the sense that it is already mentioned or at least it is a given information to the hearer.

Secondly, we focus on the unnatural sentences as the ‘X wa Y wa P’ construction. We discuss this issue by comparing ‘X wa Y ga/o P’ and ‘X wa Y wa P’.

There are some sentences which are feasible as ‘X wa Y ga/o P’ but infeasible as ‘X wa Y wa P’. The ‘X wa Y ga/o P’ construction can be formed as long as X is a subject or an upper concept of Y. Yet, in order to formulate the ‘X wa Y wa P’ construction, ‘Y wa’ must be the discourse topic as we discussed above or has to carry the contrastive meaning (Type 4). For the latter case, the following two conditions must be cleared.

1. Y has the comparative group.

2. The other comparative members have the state/event "NEG-P".

The second condition requires the implication that ‘Y2, Y3, …wa NEG-P’ has the equivalent

"Informational value" or speaker’s "assertion" to ‘Y1 wa P’. If there is no such an implication, nor Y is not the discourse topic, the ‘X wa Y wa P’ construction cannot be formed.

Close Window

Barbara F. Kelly, University of California at Santa Barbara

Ought it be considered grammaticization

This paper investigates whether ought is grammaticizing toward full modal status in two different dialects of English. It examines the current use of ought and traces changes in its use from the original source meaning as an independent verb to current auxiliary uses.

A substantial amount of research has been carried out to examine the grammaticization of modal auxiliaries in English (Coates 1983, Svartvik and Wright 1977). The development of a class of modals from a class of full verbs has occurred along a cline of grammaticization. At present ought is considered to be an auxiliary verb and it has been classed as a quasi-modal (Harris 1986, Krug 2000, Myhill 1997, Nordlinger and Traugott 1997).

Data for this study is from three databases of spoken American and British English; American English data comes from the Corpus of Spoken American English (DuBois 2000), and British English data comes from the London Lund (LL) corpus. Adult data from the CHILDES database (MacWhinney 1991) was also used. These databases were investigated to examine frequency and types of use of ought in spoken English discourse.

Ought has undergone morphosyntactic reanalysis in use from an independent verb to an auxiliary. For example, in Old English (OE) and Middle English (ME) ought was used as the main verb in the sentence, e.g. The Devil sure ought me a mishchief (1652, Brome OED). By Early New English (ENE) ought shared morphosyntactic features of modals, including question formation by inversion, negation by addition of a negative (n't/not), and absence of –s inflection on third person forms.

When we examine use of ought in Present Day English (PDE) we find evidence that ought is being used consistently in two different ways in British English, one of which is not evident in American English. This is illustrated in Table 1 below, which indicates the breakdown of overall frequency of use of ought across the database.
  British English American English RAW FIG TOTAL
ought + to 
13 (46%)
8 (89%)
ought + Æ infinitive
15 (54%)
1 (11%)
28 (100%)
9 (100%)

Table 1 Uses of ‘ought’ across British and American English databases

Table 1 above indicates that for American English uses, ought with a directly following infinitive –to is the most frequent type of use found, e.g. You ought to hear that guy play (CSAE) while for British English ought with no following infinitive (Æ ) is the most frequent form used, e.g. You ought not see him undressed (CHILDES). American English uses of ought+to totaled 89%, which is higher than the percentage of uses found in the British English corpus where ought+to accounted for 46% of uses.

This investigation spoken British and American English indicates that in British English, ought is following the same morphosyntactic historical path toward full modal status as other verbs have taken, such as should, can, and must.

Close Window

Hyun Jung Koo, Sang Myung University

On the Semantic Network and Grammaticalization of ‘Giving’ Verbs in Korean

Verbs of transference, i.e. denoting movement of an object between people, are among the words of high token frequency crosslinguistically. These verbs tend to form relational antonymic pairs as give-receive in English. In Korean, the ‘giving’ verbs exhibit higher complexity as compared with the ‘receiving’ verbs. This paper investigates three verbs of ‘giving’ in Korean, i.e. cwuta, tulita and talta, as to their semantic network of lexical and grammatical meanings. With reference to their lexical meanings, this paper describes division of labor performed by these verbs of ‘giving’. With reference to their grammatical meanings, this paper looks at their paths of grammaticalization, exploring how their meanings changed and how they interacted en route.

An idiosyncracy of the Korean language is that it has a three-way lexicalization pattern of ‘giving’, with their respective lexical meanings as in (1).

  1. a. cwuta: ‘to give’
b. tulita: ‘to give to a respectable person’

c. talta: ‘to give to the speaker or sentential subject’

As shown above, cwuta is the most basic word. It may be additionally marked with honorific marker –si-, resulting in cwusita to mean ‘to give (by a respectable giver).’ However, this honorific-marked cwusita may not be used by the speaker referring to an action of giving by the speaker him/herself, however unquestionably superior the speaker’s relative socio-cultural status may be to the receiver. This is in line with the general rule that speaker should not honorify him/herself.

Tulita, on the other hand, contains honorification of the receiver as a part of its semantics. It may be further marked with morphological honorification marker –si- to derive tulisita, which honorifies both the giver and the receiver.

Talta is peculiar in that it is not considered as an autonomous verb because its use is highly constrained both syntactically and semantically, as evidenced in the fact that most contemporary dictionaries do not include it in their entries. It is used in subordinate clauses where the sentential subject is the receiver, in which case the use of this verb is obligatory. It occurs as tallako - a form affixed with a complementizer. Its use elsewhere is in imperatives at a certain formality level where the speaker is the receiver, in its inflected form tao. In imperatives at different levels of formality, the basic verb cwuta is used instead.

These three verbs were grammaticalized in serial verb constructions, where they occurred in the second verb slot. Through reanalysis of the V1-V2 construction into V-Aux construction, the categorial status of these verbs of giving were relegated from the primary to secondary categories (cf. decategorialization), and through subjectification and semantic bleaching they acquired new grammatical meanings of benefaction and of mere direction.

A survey of historical data reveals that talta was relatively productive until Middle Korean. However, from the 17th century it yielded to cwuta. For example, a comparison of two parallel texts of an identical text, of 1518 and 1670 versions respectively, shows that the earlier text shows 8 occurrences of talta in imperatives and subordinate clauses; while the later text shows only 2 occurrences of talta and the other 6 occurrences were all replaced by cwuta – an example par excellence of functional specialization widely attested in grammaticalization studies. This specialization process is ongoing but talta is most resistant in subordinate clause usage as is manifested in that this is the only robust use in contemporary Korean.

Close Window

Alla Kushniryk, The University of Tennessee, Knoxville (visiting scholar from Chernivtsi Y.Fedkovych National University, Ukraine)

Intensifying Particles in Modern English

In Linguistics the term "particle" has several meanings, very often particle is considered to be a part of speech which gives modal or emotional emphasis to other words or group of words or clauses (Kaushanskaya 1997).

The particles could be divided into three semantic groups: connecting (also, too, either), restrictive (only, almost, merely, alone) and intensifying which include emphasising (even, just, simply), specifying (right), emotional expressive (still, quite), emphatic (then), additive (else).

According to the results of statistical investigation the intensifying particles have such peculiarities:

  1. The intensifying particles are mostly prepositional (except else and then), they stand before the words which they modify.
  2. Particles even, just, quite, simply appear in various syntactic positions. These words (except even) are never placed at the end of a sentence. Particle then occupies position at the end of a sentence.
  3. Omission of an intensifying particle doesn’t change the content of a sentence, but makes it only less emotional.
  4. The investigated particles are characterised by a great deal of syntagmatic ties and comparatively small number of paradigmatic relationships (even-simply, even-just, quite-simply). The number of syntagmatic ties of a particle depends on its positional variability.
  5. Particles even and just are the most frequently used in comparison with the other particles of the group.
  6. In the sentence the particles just, quite, simply, right, still, else occupy contact position (when the particle is placed just before the word it intensifies). Then is only used in determinant position (when the particle intensifies the whole sentence), particle even can be found in both positions.
  7. The intensifying particles constitutes more than 50% of all particles in Modern English.
The statistical investigation of the intensifying particles contributes modern linguistics with new scientific data which could be used in teaching English to speakers of other languages.
Close Window


Andrew LaVelle, University of New Mexico

The Semiotics of Time in Language: Temporal Iconicity in English Irregular Verbs

The objective of this paper will be to explore temporal iconicity in language from a Peircean semiotic standpoint, specifically focusing on the iconic mapping of the conceptual linearization of time in the present-past tense oppositions of English irregular verbs.

It is a well-known fact that a high percentage of the world’s languages, including English, use sound iconic expressions for proximal and distal forms (Woodworth 1991). The high front vowel [i] is the overwhelming favorite for signifying closeness and the low back vowel [a] for farness. Less known, however, is that, based on this paradigm, English (like its Germanic cousins) extends its phonological synesthesia to the more abstract semantic category of time as manifested in irregular verb forms which use a front vowel to express the present tense and a mid or back vowel to express the past tense (Tanz 1971). Examples are take ~ took; dig ~ dug; stand ~ stood; get ~ got; tell ~ told; teach ~ taught; slay ~ slew; speak ~ spoke. Of the approximately one hundred and fifty English irregular verbs that have vocalic alternation between their present and preterit forms, ninety percent accord with this temporal ablaut rule.

Based on Charles Peirce’s (1931-58: vol.2§277) division of iconicity into images, diagrams, and metaphors, which correspond to his three ontological categories of firstness, secondness, and thirdness, respectively, it will be argued that temporal iconicity in English irregular verbs is both a diagrammatic and metaphoric representation of the linear conceptualization of time: an articulatory diagram of proximity/distance is created which in turn evokes the underlying metaphor TIME IS A SPATIAL LINEARIZATION.

Close Window

Silke Lambert, Heinrich-Heine-Universitaet Duesseldorf Duesseldorf, Germany

What’s so special about instruments?

Whether a language possesses an instrumental case or not, instruments play a particular role cross-linguistically insofar as they can appear as structural arguments (subjects or direct objects) of verbs. This is illustrated by example (1) (see below) for English and discussed, among others, in Croft (1991). A case of special interest is Turkish, a language in which the instrumental is not recognized as a proper case because it is marked by an enclitic morpheme, whereas the ‘true’ case markers are suffixes. However, under a structural perspective, instruments display the same behavior as case-marked NPs: an instrumental NP can function as the head of a non-finite, gapping relative construction—like all case-marked NPs, but unlike all other oblique-marked phrases (cf. examples (2) and (3)). What is striking, however, is that this ‘case-like’ behavior varies with the function of the instrumental morpheme. Apart from the ‘pure’ instrumental function, it can mark comitative and manner relations, but the latter are generally less acceptable in the structures that I take as evidence for case properties (cf. example (4)).

In my talk I will demonstrate the case-like behavior of the Turkish instrumental by means of some more data and show in how far it is constrained in the comitative and manner reading. Drawing on Stolz’s (1996) cross-linguistic survey of ‘with-relations’ (Stolz’s terminology), I will discuss that a process of grammaticalization may be assumed to take place here—which, however, affects the instrumental function of the morpheme in question to a greater extent than its other functions. This can be explained by the conceptual salience of instruments in human actions and, consequently, their frequency in speech, as well as the fact that frequency is an important factor driving grammaticalization (cf. Haspelmath 2002).

An attempt to formalize the phenomenon will be presented in the framework of Lexical Decomposition Grammar (LDG; Wunderlich 1997 et passim). I assume that instruments, but no other adverbial adjuncts, can be included in verbs’ lexical entries as designated, or preferred (after Haig 1998), participants. That some kind of lexical knowledge is relevant here is shown for Turkish by the fact that, even in potentially ambiguous relative constructions, relativized instrumentals are recognized and interpreted quite unequivocally by the speakers (and the same ease of interpretation can be observed in other languages where instruments can occur in structural case configurations, as in (1)). The lexical representation allows instruments to appear in structural case positions, even if the instrumental morpheme itself is not a true case marker, as is the case in Turkish.

Close Window

Lorraine Leeson Centre for Deaf Studies, University of Dublin, Trinity College

John I. Saeed Centre for Language and Communication Studies, University of Dublin, Trinity College

Windowing of Attention in Simultaneous Constructions in Irish Sign Language (ISL)

An important characteristic of sign languages is the potential for signers to utilise two primary articulators, the hands, to express relations that occur simultaneously or quasi-simultaneously (Miller 1994, Leeson 2001). For example, locative relations are frequently encoded in Irish Sign Language (ISL) using simultaneous constructions. Other constructions that can be expressed using simultaneous articulation include topic-comment relations, articulating a lexical item on one hand while the second hand articulates a series of other signs, and placing a list of items articulated on one hand in relation to the articulation of enumeration elements on the other hand. While all sign languages have the potential to utilise simultaneous constructions in a broad range of situations, some sign languages seem to demand more regularised use of such constructions more than others: for example, Saeed et al. (2000) suggest that British Sign Language encodes simultanaeity across a range of transitive events much more frequently than Irish Sign Language, which typically uses simultaneous events to encode locative relations.

Leeson (2001) suggests that such simultaneous structures in ISL can be accounted for as occurrences of "windowing of attention" (Talmy 1996), where foregrounded elements are typically expressed on the dominant hand, while backgrounded elements are expressed on the non-dominant hand. An animacy hierarchy also seems to apply to the distribution of elements on the dominant and non-dominant hands in ISL.

In this paper we explore the windowing of attention in simultaneous constructions in ISL. We also raise some questions regarding how these structures might be accounted for in a role and reference grammar (RRG) framework (Van Valin and La Polla 1997).

Close Window

Steve Legrand and Pasi Tyrväinen, University of Jyväskylä, Finland

Connecting lexical entries to distributed real-world knowledge

Real-world knowledge is made up of myriads of everyday facts and their relationships. Without this knowledge it is practically impossible to make sense of the world when we communicate with each other. However - unlike syntax, pragmatics, and semantics - real-world knowledge is not linguistic knowledge. For this reason, it has proved problematic for linguists attempting to incorporate the required knowledge for language use into lexical entries.

Various attempts to create real-world databases have been made and are underway to address this and other problems. Lenat’s (1995) CYC, a massive database of real-world knowledge under painstaking construction for the last two decades, has been criticised by Yuret (1997) among others as being too explicit in its representation of knowledge in a single uniform framework and with deduction as its main inference engine. Locke (1990) warns about the dangers of creating systems that are accessible only to experts. To avoid the threat of ‘ontological imperialism’, the Semantic Web with its distributed ontologies and technologies seems a better alternative. XML coded and RDF-Schema based knowledge representation languages such as OIL and DAML-OIL and their future extensions with their increasing inference capabilities are suitable for domain specific, distributed ontology representation of real-world knowledge. If larger centralized database is eventually needed, then SUMO (Suggested Upper Merged Ontology) to unify disparate Semantic Web ontologies as discussed in (Pease et al 2002) could be used. Eventually, though, interfacing or even merging between CYC and SUMO might take place.

This paper will investigate the possibility of connecting distributed real-world ontologies in the Semantic Web to linguistic knowledge (syntactic-pragmatic-semantic) without corrupting the underlying aims of HPSG (Head-driven Phrase Structure Grammar) of Pollard and Sag (1994). RDF based real-world knowledge in the Semantic Web and elsewhere is divided into many distinct domain ontologies. After the knowledge about the domain is acquired by statistical free text parsing or by other means, the HPSG lexical entry’s CONTEXT|BACKGROUND attribute will be made to point to the domain ontology in the Semantic Web and to structure-share it with HPSG’s other components. The pointed domain ontology will be retrieved and aligned with semantic upper ontology integrated in the lexical entry and based on Pustejovsky’s (1995) qualia structure as exemplified by Verspoor (1997) for nominals and on Jackendoff’s (1983, 1990) lexical conceptual structures as used by Davis (1995) for verbs. As the result, the real-world knowledge, together with semantics, syntax and pragmatics can be integrated to constrain the structure-shared lexical entries. It should be kept in mind that although the knowledge about the applicable domain is determined by pragmatics in this case, knowledge contained by the domain ontology itself is real-world knowledge.

The main motivation behind this research is to improve the accuracy of linguistic parsers to benefit linguistic applications used in translation and language learning and other tasks, which use parsers for disambiguation. Current parsing applications might seem adequate for these purposes having reached accuracies close to 100 per cent. However, a word-based disambiguation error rate as small as 4 % is high enough to completely change the meaning of an average-length sentence translating into a 56% per-sentence error rate (Abney 1996). Deployment of real-world knowledge together with linguistic knowledge in disambiguation will help to bridge this gap.

Close Window

Zouhair Maalej, University of Manouba, Tunis, Tunisia and Visiting Fullbrite Scholar at the University of New Mexico

Naming People in Tunisian Arabic: An Idealized Cognitive Models

Naming people has been studied referentially (Quirk et al, 1972), etymologically (Jäkel, 1999), pragmatically (Marmaridou, 1989), and
syntactico-semantically (Van Langendonck, 1999). The present paper, however, offers a cognitive semantic view of naming in Tunisian Arabic (TA) as an Idealised Cognitive Model (Lakoff, 1987; Langacker, 1991). First names are regarded as semantically motivated but unconscious strategies, describing a propositional model (Lakoff, 1987), a trigger-target image-schema (Fauconnier, 1994; Johnson, 1987), a part-whole metonymic model (Lakoff, 1987), and a motivated metaphoric mapping (Lakoff and Johnson, 1980-1999). Prototypically, first names in TA rely on a  conceptual domain either recruited from within the experience of name-givers or are the outcome of a desirable state of affairs on the part of name-givers. Such a desirable state of affairs is the product of an imaginative projection on the part of name-givers, who build this projection from within emotions, morality, beauty, piety, etc. The conceptual metaphors capitalised upon in naming reveal name-givers' perception of males and females in TA, with blatant bias to the former. The result is a cultural model of naming, whereby conceptual metaphors interface with derivational morphology.

Close Window

Candace Maher, University of New Mexico

Theme, Thought, and Theory: Jicarilla Apache Classificatory Verb Stems

In this paper, I examine Jicarilla Apache classificatory verb stems, a type of verbal classifier which is employed to classify things in the world. Stative classificatory verb stems (CVSs) incorporate a noun’s physical attributes and spatial location. Several physical attributes that categorization is based on are: shape, dimension, animacy, liquidity, configuration, and number. Spatial location refers to the position (e.g., sitting, resting, in position) or existence of a noun (subject). Motion CVSs describe motion type of a noun (object): 1) contact with an agent, 2) no contact with an agent, or 3) some kind of propulsion concomitantly encoding physical properties. These types of motion are based on Navajo Hoijer (1963) and Young & Morgan (1980). Semantic parameters and domains are explored in the Jicarilla Apache classification system. I explore the underpinnings of the Jicarilla Apache conceptual system and the use of these very rare CVSs. Jicarilla Apache is a polysynthetic languagein which words are created from a verb with multiple prefixes. The CVS is determined by noun qualities as well as motion type (stative or active).

Close Window

David R. Margolin, University of New Mexico-Valencia

The Joy of Tawahka: An Overview of the Phonology, Morphology, and Syntax of a Misumalpan Language

 The extant Misumalpan languages (Miskitu, Mayangna/Northern Sumu, and Ulwa/Southern Sumu) are spoken in Central America, in Honduras and Nicaragua. All together their speakers number around 100,000: there are approximately 80,000 speakers of Miskitu, 10,000 speakers of Mayangna/Northern Sumu, and less than 1,000 speakers of Ulwa/Southern Sumu. The two Sumu languages are in imminent danger of disappearing, largely due to the overwhelming influence of the Miskitu ethnic group and its language, which is the dominant language in this tropical rainforest region. Virtually all Sumu speakers also speak Miskitu, in addition to Spanish and/or English; virtually no Miskitu speaks a Sumu language.

All these languages are of the head-final type and are astonishingly similar in their grammars, so much so that describing any one of them describes perhaps 80 to 90% of any other. Ken Hale and his associates at MIT (Hale 1991, 1994; Benedicto and Hale 2000) suggest that this situation results from long-term bilingualism in the area, which is very plausible given the history of intense interaction between the Miskitus, a group that has expanded territorially and numerically since the British furnished a small coastal enclave with muskets in the seventeenth century, and the Sumu speakers.

Tawahka, the Northern Sumu variant spoken in Honduras (with which I have been working for the last year as the result of a Fulbright grant) exhibits a number of interesting grammatical features, which will be discussed in the presentation: a phonological system with three vowels and a reduced consonant inventory; weight-dependent affixation, resulting in affixes appearing both as suffixes and as infixes; extensive verbal morphology, with semantically intriguing formal similarities between third-person and first-person-plural-inclusive forms; a formal continuum between "normal" verbs, stative verbs, and adjectives; possession of nouns is marked on the noun, not the possesor; and a comprehensive system of switch reference used in sequential verb clause chaining.

Close Window

Elisa Maroney, University of New Mexico

Aspect in American Sign Language: Expression of Signer Perspective

Aspect is the linguistic expression of a signer’s or a speaker’s perspective on the internal temporal structure of an event. How do American Sign Language (ASL) users perceive the internal temporal structure of an event? How is this perspective expressed inguistically in ASL? Much of the research that has been done on aspect in ASL was done before substantial cross-linguistic work on aspect had been done on spoken languages. While the existing research on aspect in ASL does provide valuable formal and semantic descriptions, the labels and descriptions that were applied will be re-evaluated in light of the cross-linguistic study that has been done on aspect in spoken languages, as well as the research that has emerged in other signed languages.

The goal of this presentation is to report on findings about the aspectual patterns of ASL elicited by using a structured questionnaire adapted from Dahl (1985). The most common aspects found in spoken languages were included to determine whether or not these categories exist in ASL. The inflectional and derivational properties will be discussed to determine where the aspectual categories described in ASL fit into the inflectional-derivational continuum. This study of ASL will provide an opportunity for comparisons with the aspectual systems of spoken languages, as well as a base from which a cross-linguistic study can be done on aspectual categories across signed languages. The results of this study of aspect in ASL will be compared to Dahl’s results of the 64 spoken languages in his study.

Close Window

Kyoko Masuda, University of Arizona and Carlee Arnett, University of California, Davis

Dative Subjects in German and Japanese

This paper will compare the semantic characteristics of participants that occur with the dative case in German and certain participants that occur with the postposition ni in Japanese. The dative case in German and ni when used as a dative marker tend to mark experiencers. The use of the dative case to mark an experiencer is not uncommon cross-linguistically, but in Japanese the experiencer participant is considered to be part of a double subject construction (Kuno 1973; Kabata & Rice 1997; Shibantani 1999; Kumashiro 2000) and in German the subject-like properties of the participant are rarely considered.

Using Langacker’s (1991) Action Chain Model, we compare German and Japanese dative subjects. In both German and Japanese, the experiencer participants are active participants in the event, tend to be animate, definite and the topic of the sentence. Consider the following examples (Kabata & Rice 1997:114):

  1. Kare ni wa musako no gokaku [sic]ga totemo uresikat-ta.



    He DAT TOP son GEN success S very happy-PAST

    (As for him,) he was very pleased with his son’s success.

  3. Mir gefaellt deine Bluse.
Me-DAT please-PRES your blouse.

I like your blouse.

As example (1) shows, in Japanese, these relations are marked by topic marker wa and subject marker ga. The semantic characteristics are often overtly marked in Japanese, but there is less overt marking in German.

The focus in this paper will be on the written language. For both Japanese and German, the data comes from current newspapers, periodicals and novels. These text types are representative of the standard language.

In this paper, we describe the event structure and other semantic properties of the Japanese double subject construction with dative experiencers and show that the same type of construction is found in German. Based on these semantic similarities, we can argue for a double subject construction as well as for experiencer subjects in German.

Close Window

Mami Osaki McCraw , University of New Mexico

Evaluation in Japanese Oral Narrative

This study analyzes Japanese oral narrative. According to Labov (1972), narrative has six components: abstract, orientation, complicating action, result/ resolution, coda, and evaluation. Evaluation is one of the most important elements in narrative structure since it tells the narrator’s point. I examine how the Japanese narrator expresses his or her point in oral personal narrative. The subjects are two Japanese speakers in middle-class (one male in his 20s and one female in her early 30s). The data were taken from approximate 90-minute informal conversations between two friends. Each of the actual narrative portions lasts about five minutes. From the analysis of these two data, it turned out that the Japanese narrator uses not only external evaluation but also internal evaluation. The narrator makes evaluative points using sentence final particles, interacts with his or her hearer in order to give a clearer picture of the narrative world, and expects the hearer to infer the evaluation from the given background information. The implicit evaluative strategies used by the subjects concur with the findings of a previous study by Okazaki (1993), which she calls "listener-dependent strategies" (p. 93). In Japanese culture, which values humbleness and prefers the indirect communicative style, it appears reasonable that the narrator rather embeds evaluative comments than uses external evaluation. In future research, it is recommended that more data of Japanese narrative be compared intra-culturally (e.g., gender, age, region, social class) and cross-culturally/ cross-linguistically (e.g., English, Korean, Chinese) in order to investigate similarities and differences within and across cultures.

Close Window

Ana Aurora Medina Murillo, University of New Mexico

From compounding to affixation

It is common knowledge that many affixes we find in the languages of the world originated from members of a compound word. This happens because of the "demotivation" process that such a compound element may undergo. The basic criterion to distinguish a compound element from an affix is that the former also occurs as a free word, while the latter is always attached to another element and has a simpler form. Similarly, the compound element has more lexical meaning, while the affix has a more general one. We can identify, however, a number of lexical items which are in a transitional stage from compound element to an affix; the so-called affixoids (Olsen, 2000). In the present paper I will illustrate the diachronic development of some lexical elements of Guarijio, which we may consider affixoids. For that purpose I will consider the frequency with which the element appears in compounds, the deviance of meaning of such element from the basic meaning of the free form, and the syntactic characteristics of the resulting form.

 1. a. simi- 'to go'

b. ca-/ce- ‘to say’ + -si ‘go’ = cesi- ‘to come saying’

c. wehpá- ‘to beat’ + -si = wehpisí- ‘to start beating’

2. a. puha- ‘to take away’

b. ma’sá ‘feather’ + -pu = ma’sapú- ‘to pluck’

c. te’ki- ‘to go down’ + -pu = te’kipú- ‘to put down’

Close Window

Brenda Nicodemus and James MacFarlane, University of New Mexico

Frequency & Acceptability in ASL Phonotactics

In a study of phonotactic constraints in ASL, the question of whether linguistic experience is instrumental in the mental representation of language is explored.

A nonce experiment designed to test subjects’ ability to rank nonce signs with respect to "goodness" was performed. Subjects (n=24) were presented with nonce signs in which the handshapes were replaced with frequent handshapes, infrequent handshapes, and non-occurring (impossible) handshapes in ASL. The results indicate that subjects do maintain statistically significant preferences for the nonce signs that employ the most frequent material in the language. It is suggested that a signer’s experience with language, as opposed to an abstract set of categorical rules, is a key factor in the mental representation of language.

Close Window

Loretta M. O'Connor, Max Planck Institute for Psycholinguistics, Nijmegen & University of California, Santa Barbara

Manner-salience and state change in Lowland Chontal of Oaxaca

Lowland Chontal of Oaxaca, an endangered and unclassified indigenous language of Mexico, encodes the manner of a change of location or change of state both as an independent predicate and as the initial element in a compound stem. This paper addresses two questions that arise from the examination of both types of constructions. First, do 'motion event' predicates with compound stems entail a route or path (Jackendoff 1983; Talmy 1985, 2000), or do they describe the resulting physical location or disposition achieved by the semantic Theme, leaving the route or path element to pragmatic inference? I will argue that elements that behave like directionals, a typical feature of Mesoamerican languages, often depict an endstate in Chontal. The second question focuses on the distribution of the two types of constructions. Evidence will show a general pattern in which state and location changes performed by humans, especially but not exclusively with the hands, are encoded with compound stem predicates, while multiple clauses are required to express events in which human involvement is less immediate. The syntactic patterns of multi-clause expressions provide synchronic clues to probable diachronic processes and suggest the role of discourse frequency in the evolution of compound stems (Bybee & Hopper 2001). Furthermore, examples from natural discourse show that compound stem predicates depict manner-salient events that occur at locations situated by previous motion events. Thus, the conclusions also contribute a perspective from an underdescribed language to our understanding of manner-salience (Slobin 2002) in motion and state change events. The data used in this paper are from elicitation, elicited narrative, and natural discourse, collected in primary fieldwork in and around San Pedro Huamelula, Oaxaca.

Lowland Chontal is a verb-initial, head-marking language with variable word order, no case marking, one preposition, and a complex system of aspect but no tense. The morphological type is agglutinative, with mostly prefixes on nouns and mostly suffixes on verbs. The major person-marking paradigm is an agentive system motivated by the control or agency of the participant.

The language also has a productive pattern of complex predicate formation in which two or more verbal elements combine to form a stem expressing change of location, position, or state. In these predicates, at least one directional element (from a closed set of about a dozen members) follows an initial element of manner of motion, manner of contact, or manner of spatial configuration to form a compound stem. Semantics of directionals include familiar concepts of 'up, down, across, into'. For example, the morpheme ñi- 'across' is used in predicates of 'crossing' and 'receiving', as in (1), and it also encodes a resulting state of 'apart', (2).

(1) jway-ñi -pa el pana' le -ñi -p -ola' laka'no'

jump-across-pfv.sg det river human-across-pfv-3p.pat woman

'He jumped across the river.' 'The woman received them (in her home).'

(2) jas -pa lich'ale jas -ñi -pa lich'ale

tear-pfv.sg cloth tear-across-pfv.sg cloth

'She tore the cloth.' 'She tore the cloth apart.'

However, notions such as 'swim across' cannot be expressed with a compound stem; for these, a multi-clause construction is used, as in (3).

(3) xnoy-yuy ki -ñi -pa el pana'

swim-dur.sg vTraj:straight-across-pfv.sg det river

'He swam across the river.'

The order of the two predicates in (3) suggests a process in which frequently occurring manner + motion/endstate combinations grammaticized into compound stems.

Close Window

Winnie W. F. Or, University of Texas at Arlington

Trading Places: Discourse and Identity in Chinese Local Markets

A number of studies of business negotiation have focused on structural characteristics of such activities. Some researchers interested in whole encounters have attempted to identify different phases of business meetings or service encounters involving buyers and sellers (e.g., Tsuda 1984); others have focused on specific utterance sequences claimed to be characteristic of negotiating encounters, e.g., the bargaining sequence (Maynard 1984, Hester 1990, Firth 1991). There is indeed a certain appeal to analyzing highly predictable and orderly interactions between buyers and sellers in structural terms, i.e., breaking the discourse down into its constituent parts, such as chronological phases or utterance sequences. However, the patterns or sequences that are readily identifiable in buying and selling negotiations in one setting may not be such in another. In fact, the premise that people engage in negotiating activity by following some pre-determined structures, as some of these structural studies seem to suggest, is fundamentally flawed.

I argue that a social/critical perspective to discourse (e.g., Van Dijk 1997, Fairclough 1992, 1995) is a much more informative approach to analyzing what goes on in a negotiation encounter. Such an approach takes into account not only the sequential context where an utterance occurs, but also the wider social structures and practices, including cultural values, beliefs, and knowledge of members of a community, within which discursive activities take place. Emphasis is placed on the constitutive function of discourse: producing discourse is part of the wider processes of producing social identities, social relations, and social life in general. Discursive practices are hence social practices that may either reproduce, or contest and change, social and cultural structures.

With this perspective, I examine mundane and everyday buying and selling encounters in the traditional setting of local markets in southern China, where interactions between vendors and shoppers are highly routinized. The regularity and predictability of the interaction in these encounters are analyzed as a set of discursive practices that participants adopt which reproduces (1) the social identities of traditional buyers and sellers, and (2) the conventional social relation that obtains between them. Regarding (1), the tightly circumscribed social identities of the traditional buyer and seller are clearly demonstrated on the action level where the interaction is dominated by bargaining over price and quantity. Regarding (2), the strong asymmetry in the social relation between the participants is manifested on the exchange level where sellers are responsible for a much larger part of the conversational labor, are constrained by more obligations, and enjoy less latitude in talk.

The analysis of discursive practices in buying and selling activities in a traditional, prototypical setting like the local market is significant in that it contributes to the understanding of the more ambivalent and contradictory practices observed in negotiation encounters in other less traditional settings such as corporate boardrooms.

Close Window


Jung-ran Park, Indiana State University ; University of Hawaii

Politeness and Solidarity in Korean Address Forms

The social structure of any society is closely reflected in forms of address. In Korean, the use of names and the second person pronoun (English ‘you’) as address forms is restricted to certain contexts, such as asymmetrical dyads and children’s language (Brown and Gilman 1960). This implies a tendency toward avoidance in addressing intimates as well as unknown individuals in the environment of public spaces. This result is a gap in the address system.

The Korean pragmatic marker ceki(yo) ‘excuse me/well’ evinces multifunctional aspects (Park 2001). One function is addressing the unknown individual. As an address form, ceki(yo) partially fills the aforementioned gap, especially between strangers whose relationship is based on degrees of social distance. An address form complementary to the function of ceki(yo), ipwa(yo) ‘hey/look’, fills the address form gap for close friends, family members, etc., whose relationships are based on solidarity.

The following is an analysis of morphemes that are agglutinated onto these two address forms:

This paper aims, through an examination of the Korean deictic system, to determine the underlying characteristics and motivation for the complementary functions of the above two Korean address forms on the basis of the dimension of social distance in informal social contexts. Another aim is to find usages manifesting strategic politeness (Brown and Levinson 1987) involving the address form ceki (yo) through examining characteristics of its source meaning: i.e. indirectness and neutrality. For the study, I analyzed 200 tokens of the two address forms in naturally occurring conversation, including data from contemporary Korean television dramas.

The address forms ceki(yo) and ipwa(yo) evolved from distal and proximal deictic elements, respectively, serving complementary functions based on characteristics of the source meanings. The two address forms underwent semantic and pragmatic change through the grammaticalization processes (Park 2002). Physical distance, the original determinant of the proximal and distal deictics was pragmatically transformed into psychological distance for social interactions through metaphorical extension of the original meaning. Specifically, ipwayo is derived from the proximal deictic element marking solidarity and directness, whereas cekiyo is derived from the distal deictic element marking distance and indirectness.

The distal deictic ce is unmarked; that is, it is employed when referring to a referent distant from both speaker and hearer. Unlike the proximal deictic i, which is speaker-perspective oriented, ce provides a basis for expressing the speaker’s polite attitude owing to its indirect and neutral characteristics. This furthers the function of ceki(yo) as an address form for unfamiliar interlocutors while allowing for the strategic expression of politeness.

Close Window

Milena Petrova, University of Southern California

Bulgarian "some" indefinites

This paper investigates the distribution and semantic properties of the Bulgarian indefinite phrases "njakoj N" (some-who/which N) and "njakakuv N" (some-what N), both of which are glossed as the English "some N". My main proposal derives from the idea that both determiners inherit the semantics of their constituent wh- morphemes; namely, I argue that "njakoj N" is a covert partitive and its restriction is D-linked, which explains its ungrammaticality in existential "there" insertion contexts (1). It is not suggested, however, that "njakoj N" has a covert definite in its structure as it can be overtly embedded in a partitive phrase (2). Furthermore, the unrestricted form "njakoj" (some-who) is perfectly acceptable in (1), presumably because it is not D-linked.

Crucially, the distribution of "njakoj N" in Indicative Present and Past Tense sentences is extremely restricted (3-4), and indeed ungrammatical on most occasions. On the other hand, both "njakoj N" and "njakakuv N" are licensed in modal and quantified contexts (5-6). However, only "njakakuv N" can take wide scope. My proposal is that this peculiar property of "njakoj N" is due to a clash of felicity conditions between its use and the use of the Indicative Present and Past Tenses. My analysis of the compositional meaning of "njakoj N" (some-who/which N) predicts that the phrase is acceptable in the same contexts where its wh- counterpart "koj N" (who/which N) is acceptable (cf. 1); namely, the contexts in which the speaker is not able to exhibit the witness set N. At the same time, The Indicative Present and Past Tenses in Bulgarian require that the speaker has witnessed the event or has been sufficiently involved in it so that she is able to form a vivid representation. If the speaker has a vivid representation of the event, however, she will also be able to exhibit the individual that will make the proposition true; thus, the conflict marks the use of "njakoj N" as infelicitous. When the DP contains only the unrestricted form "njakoj", the condition is vacuously satisfied as there are no witnesses at all (7).

This analysis predicts that if there is an event of which the speaker has a vivid representation but for some reason or other is unable to exhibit the witness, "njakoj N" will be allowed, which is indeed the case. Furtheremore, "njakoj N" is freely licensed with the Perfect Tense (8), which Izvorski (1997) argues is an epistemic modal, thus confirming the hypothesis. Moreover, I argue that "njakoj N" can take wide scope in (9) because the matrix clause tense is a modal.

As far as "njakakuv N" (some-what N) is concerned, I suggest that it is also licensed when the corresponding wh- question "what", or more precisely "what-kind-of" is licensed. Thus, "njakakuv N" is interpreted as a generic predicate, indicating that the speaker is not familiar enough with the referent to be able to apply a more specific or descriptive one. Thus, when both "njakoj N" and "njakakuv N" are licensed (as the narrow scope reading of 5-6) the difference in interpretation is very subtle: whereas the referent of "njakoj N" must be picked out of a contextually salient set of individuals, the referent of "njakakuv N" is completely unspecified. The distinction becomes clearer in examples like (10-11). Suppose that I leave 3 kids at home with 3 biscuits on the table; when, on my return, I see that all the biscuits are gone I can utter (10) with "njakoj N" because there was a contextually salient set. "Njakakuv N" is infelicitous because, being familiar with the biscuits, I cannot use a generic predicate. In (11), on the other hand, the reverse is true: suppose that I see my kids going out of the candy store, each carrying a candy. "Njakoj N" is not licensed in this context because I do not have a contextually salient set in mind; however, they each carry "some-kind-of" candy, so I can use "najakakuv N".

Finally, neither determiner has a cardinality meaning, which is why neither is licensed in contexts where that is the only possible interpretation (12-13). Instead, speakers must use a third determiner, "njakolko N" (some-how-many).

Close Window

Seongha Rhee, Hankuk University of Foreign Studies

Grammaticalization of malta 'stop' in Korean

In contemporary Korean, malta, a verb denoting cessation or 'stop', displays peculiarities in several aspects. For instance, it is rarely used as a main verb; instead it is used predominantly as an auxiliary verb. In some cases when used as an auxiliary, this verb, despite its transparent verbal status, is classified as an auxiliary adjective, a category so unusual that many linguists and grammarians are reluctant to recognize it as a grammatical category in Korean. In addition, there are numerous fossilized forms containing this verb, of which the exact meaning, however, is hard to define. This paper explores these phenomena from a grammaticalization perspective.

A historical survey of the usage of malta reveals characteristic developments associated with grammaticalization, such as decategorialization, divergence, semantic generalization, functional enrichment, gradualness of grammaticalization, etc. For example, a comparative survey of four translations of an identical text, the 16th, 17th, 18th century and Modern Korean versions, shows that in the 16th century the verb was actively used as a main verb (28 tokens), whereas such use is rapidly falling into disuse in a modern translation (3 tokens). The gradualness of this process is well illustrated in the token frequencies in these four texts: 28-24-19-3. On the other hand, its use as an auxiliary verb is on the increase, albeit with a minor fluctuation en route, as shown by 52% of all uses in the 16th century version and 95% in the Modern Korean version.

As a grammaticalized item (combined with a non-finite marker -ci), its primary function is to encode 'prohibition'. However, this prohibitive marker underwent semantic attrition, whereby the semantic component 'volition' as to cessation becomes lost, thus enabling the auxiliary verb to co-occur in non-volitional context e.g. with psych-verbs (combined with a non-finite marker -e). This possibility triggered creation of expressions that can take inanimate subjects. Furthermore, since the original semantics of the verb relates to cessation, the verb also created other auxiliary uses for emphasis..

The paths it has taken in history show that malta is a paradigm example of verbal grammaticalization.

Close Window

Cinzia Russi, University of Washington

The allomorphy of the Italian masculine definite article: A ‘language-use’ based account

In Modern Standard Italian (MSI) the masculine definite article shows the following allomorphs: singular il, lo and l (l’ orthographically); plural i, gli [´i]. Their distribution is illustrated in (1):

1. Italian masculine definite articles

    1. il sg./i pl. occur before single consonant other than /S, ø / (/ilraÈgats:o/ ‘the boy’, /ilÈmare/ ‘the sea’); the affricates /tS, dZ/ (/ilÈtSElo/ ‘the sky’, /ilÈdZelo/ ‘the frost’); and consonant clusters of the type stop + /r, l/ (/ilÈgril:o/‘the cricket’, /ilÈpliko/‘the package’).
    2. lo sg./gli pl. occur before consonant clusters of the type /s/ + C (/lostuÈdente/ ‘the student’, /loÈsba´o/‘the mistake’); the affricates /ts, dz/ (/loÈtsio/ ‘the uncle’, /loÈdzero/ ‘the zero’); /j, S, ø/ (/loÈjdio/ ‘the iodine’, /loÈSal:e/ ‘the shawl’, /loÈøomo/ ‘the gnome’); and /ps, pn, pt, ks,kt/ (/lopsiÈkologo/ ‘the psychologist’, /lopneuÈmatiko/ ‘the tire’, / lopteroÈdat:ilo/ ‘the pterodactyl’, /loksiÈlŤfono/ ‘the xylophone’, /´ikteÈnofori/ ‘the ctenophores".
    3. l sg./gli pl. occur before vowel, diphthong (/laÈmiko/ ‘the friend’, /lauÈtista/ ‘the driver’), and the labiovelar glide /w/ in original Italian lexical items (/lÈwŤvo/‘the egg’) .
Traditionally, the allomorphy of the Italian masculine definite article has been accounted for entirely in phonological terms, i.e. the opaque distribution of the three allomorphs il, lo and l has been related to the features of the initial segment/cluster of the lexical item following the article (Mulja?i? 1971, 1974; Davis 1990; Vanelli 1992). In my opinion, however, none of such phonological accounts appears fully satisfactory, especially with respect to the following three issues: a) the input of the rules; b) the nature of the rules themselves; and c) the environment in which such rules apply. Moreover, phonological analyses often appear too dependent on quite abstract and empirically questionable principles. In contrast, a morphology based account seems more convincing and, most importantly, allows for a unified account of the opaque distribution of the allomorphs, both from a synchronic and a diachronic perspective (Russi 2001).

The main goal of this paper is to show that the usage-basedframework proposed by Bybee (2001) can be adequately applied to the analysis of the distribution of the Italian masculine definite article. Also, I would like to argue that a usage-based based account can offer new insights to this unquestionably complex phenomenon, in particular regarding the issues of individual variation and changes that can be observed in the current distribution pattern, such as the generalization of il to autochthonous (phonological) environments where lo would be expected as well as loanwords.

Close Window


Tetsuya Sato, The University of Arizona

Gendered Voicedness: Style Variations in Direct Quotations as Evidence for Gender-Indexing in Japanese Conversational Narratives

 In this paper I explore 'gendered voicedness' of direct quotation in Japanese conversational narratives, based on Bakhtin's notion of "double-voiced discourse" (1981:324). The speaker's gender identity and language use in Japanese conversation has been extensively studied (cf. Okamoto and Sato 1992; Okamoto 1994, 1996; Matsumoto 2001). However, 'gendered' speech in direct quotation has not been paid much attention so far, to my knowledge. Thus, I focus on this issue in the present study. Through the examination of direct quotation in audiotaped Japanese conversations, I argue that 'gendered' language ideologies are manifest in direct quotation, where the quoter performs the quotee's 'female/male' identity through talk. That is, by directly quoting other's speech, the quoter performs the socio-cultural identity of the quotee, including his/her gender identity as an integral attribute to that individual. More specifically, when the speaker incorporates other's voice (i.e. speech) in storytelling, treating it as if it is verbatim, he or she utilizes his or her metalinguistic/metapragmatic knowledge of gender-indexing linguistic resources in reproducing how the original speech was delivered. Conversely, direct quotation seems to be a reasonable place to look for an evidence for 'gendered' language use in Japanese.

Observe the following example of a direct quote. Speaker M is reporting to her friends a conversation with her mother regarding aging.

213 M: [babaa].


214 toka tte,

like QT

215 fuzakete yuu to,

teasingly say then

When I said to my mom teasingly, "old woman," or something like that,

-> 216 ... anta mo,

you also

-> 217 koo nan no yo.

this become NOM FP

"You're gonna be like this too, you know,"

218 <@ tte @>.


In lines 216 and 217, M directly quotes her mother. In the sentence-final position of this direct quotation, M uses the particle no followed by final particle yo, which is traditionally classified as a "super-feminine form" (Okamoto and Sato 1992: 481).

All the conversations were casually exchanged among friends and family members. The participants vary in terms of age and gender but all speak primarily Standard Japanese. The data for this study consists of fourteen audio-taped naturally occurring conversations in Japanese (approximately 108 minutes total). I used both formal (quotation forms) and prosodic criteria to identify all the possible direct quotes.

Close Window


Sagit Shafikov, Ufa, Bashkir University, Russia

Semantic universals in language vocabulary

This study is devoted to ascertaining semantic universals in language vocabulary, namely the vocabularies of Russian, English, French and Tartar. Until recently research into linguistic universals has primarily focused on phonology, morphology and syntax, although to analyze the lexical domain is more vital for contrasting language world-views and establishing categories common to all languages irrespective of origin and structure.

The aim of the study is to ascertain common semantic structures in identical lexical sets (semantic fields) and interpret constraints on language variation within 4 levels of analysis: 1) level of sememes, i.e. denotational lexical meanings integrated into a semantic field by one common seme, 2) level of semantemes, i.e. bunches of sememes corresponding to one lexeme, 3) level of phraseological units related to lexemes under analysis, 4) level of the inner form of lexemes which makes part of word semantic structure and refers to an image underlying it usually based on metaphor or metonymy.

The first level introduces lexical categories (sememes) identical in the contrasted languages with reference to a few semantic fields including kinship and affinity terminology, body-part terminology, animal terminology, “agent” and “physical perception”. Constraints on language variation are explained thru the hypothesis of functional convenience. The second level introduces semantic structures of polysemous lexemes in contrasted languages revealing identical associations between primary and derived sememes, for example “head”: “leader”, “donkey”: “stubborn person”, “to see”: “to understand”, “son”: “boy”(as a form of address), etc. The third level deals with semantically and structurally equivalent phrases containing identical component lexemes in the contrasted languages. Analysis shows that in each language derived sememes are potentially connected to phraseological meanings making up phrasal periphery of the corresponding sememes, for example English take in hand = Russian vzyat v ruki  = French prendre en main = Tartar khulga alu. As idioms are characterized by emotive evaluation it is interesting to correlate positive and negative evaluation in cross-linguistically identical phrases. An analysis based on idioms related to body-part-terminology proves that negative evaluation prevails and tends to be expressed by lexemes pointing to relatively less important body-parts such. The fourth level refers to the motivating structure of the lexeme which may include one or both of two motivating components, namely identifier and modifier, depending on whether the motivating structure points to genus and/or species.

Due to the full structure type, which prevails in English and Tartar the lexeme in each of these languages, has a full informative value, as opposed to Russian and French, whose lexemes have relatively less semantic predictability potential.

Close Window

Asa M. Stepak, Independent Researcher, Cum Laude graduate of New York University

Rationalism Versus Empiricism: A New Perspective

The rationalist view, popularized by the Chomskyan notion of ‘poverty of the stimulus’, is contraindicated by the empiricist viewpoint that language learning and use is governed by general cognitive principles. The rationalist viewpoint proposes that language use is under the control of a primitive cognitive functionality that can be characterized as a modular unit or organ relying on a unique set of architectural/dynamic principles. The empiricist viewpoint proposes that language use is governed by general cognitive principles that govern all cognitive behaviors and which follow a common pathway. The rationalist point of view considers cognition mechanisms as being domain specific for language whereas the empiricist point of view believes them to be non-domain specific. In support of the rationalist viewpoint, Chomskyans rely on the complexity of grammatical surface structure as proof that language learning involves learning of structural blocks that have a gestaltist nature to them in that learning of the ‘whole’ precedes learning of its individual components. In other words, grammatical structure is learned in blocks intuitively that could not be explained by an empirical learning of the component parts comprising of the structure. Further, Chomskyans claim that the grammatical structure of language is not a logical representation of components parts that could be learned—in other words that the grammatical structure itself, acquired intuitively, is irreducible.

I will present argument opposing the Chomskyan viewpoint. Relying on new concepts I have introduced, ‘Oral Metaphor Construct’ (OMC) and ‘Knowledge Inheritance’ and basic principles in Information Theory, I will argue that the visible generative grammatical structure of language is not what is actually learned but ,rather, represents a pattern (fractal) of more fundamental learning that is ‘cognitively linear’. In other words, what is actually learned could be described in terms of its component parts, thus, supporting the empiricist notion of a non-domain specific cognitive mechanism governing language use and refuting the Chomskyan viewpoint.

Close Window

Laurel Smith Stvan, University of Texas at Arlington

Why and Say: Two Discourse Markers in Depicted Conversation

Most work in the area of discourse markers (e.g. Schiffrin 1987, Heritage 1998) has focused on the commonly used utterance-initial forms oh, well, and so. This paper focuses on two other markers which occur in the same environment: why and say. While the first three occur quite frequently in naturally–occurring verbal exchanges (collected from online transcripts and the Brown corpus), the latter two do not; the bulk of their occurrences are found in the speech depicted in written or televised fiction (e.g., in one half-hour transcript from a 2001 TV interview show, there were 12 wells, 12 sos, 8 ohs, but no whysor says). The data indicate that why and say are most often used in modern prose that depicts a conversation.

When characters do use non-interrogative why, it exhibits four distinct functions: First, it is used as a way to fill a moment while ruminating, as seen in (1). Found mainly in set phrases—particularly why yes, why no, and why thank you—this why underlines what follows. In its second use, why is found as the preface to an exclamation, typified by the Three Stooges cry, "Why, I oughta…," as well as the example in (2). While the first two uses set up new material, the remaining two uses link new utterances to previous information (Green 1996). Thus, the third use of why is to indicate surprise at the self-evidentness of the previous query, as seen in (3). Fourthly, why appears as a way to initiate a consequence within the utterance of a single speaker. As seen in (4), it serves as a shortened form of why then. Say, on the other hand, is used in two related ways, one as a vocative summons (more commonly represented with hey) as seen in (5), and the other as a presentational device for setting up a suggestion, as seen in (6). This second use is similar in function to the exclamatory use of why.

Recall the distinction between the naturally spoken uses of oh, well, and so and the depicted uses of why and say. Another observation is that oh, well, and so occur frequently in contemporary English, while why and say used in the ways discussed above appear most often in depictions of conversation between non-contemporary characters in fiction. While both sets are still in use, I suggest that say is the most dated form, having had more currency in the early to mid-twentieth century ("Say, pal, can you tell me the time?"). Utterance-initial, non-interrogative why seems to be more current than say, but more dated than oh, well, and so. Because oral language is often more innovative than written forms, I suspect that when currently spoken, the particles why and say sound somewhat formal and dated. Further analysis of corpora from different eras should verify the frequency of the members of each set by speakers of different ages. Thus connotations of the spoken form and occurrence of data from different ages of corpora indicate a difference in the datedness of the forms. Nevertheless, most likely because of their positioning abilities, why and say continue to be used as a stylistic device by authors and playwrights in depicting conversations.

Close Window

Maria Szasz, University of New Mexico

Ireland’s Astounding Nineteenth Century Language Shift

Language history in Ireland, like its political and religious history, has a complex past. In this paper, my primary concern is to understand one of the most important events in Ireland’s linguistic history: why the English language began to spread rapidly across Ireland in the 1800s, picking up speed until it almost eclipsed Irish Gaelic by the end of the century.

Uncovering the causes behind such a drastic language shift demands a consideration of several factors. First, the Act of Union joining Ireland with England in 1801 greatly increased the transportation and commerce between the two countries. Following the Union, the modernization of the Irish railway system soon linked Ireland’s main English-speaking cities with the predominately Irish-speaking, rural parts of the country for the first time. This rapidly expanded the frequency and proximity of English influences across virtually all of Ireland.

Second, the creation of an Irish National School system in 1831, with its English-only instruction policy, also played a significant role in decreasing the younger generation’s use of Irish. Third, the Irish Famine, lasting from 1845-1852, took its greatest toll on the primarily Irish-speaking population, through death, emigration and the Irish community’s deliberate decision to forego Irish for English. Finally, the prominent social, political and religious opinions in the nineteenth century also heavily favored English. The circumstances behind this rise of English and the corresponding decline of Irish that I find most intriguing, and which I would like to discuss in detail, lie in two areas: the establishment of the National Schools, and how the Famine affected the Irish people’s views about their native language. Through an exploration of these two factors, I hope to explain why English rose to linguistic dominance in Ireland in only one hundred years, becoming the sole language of eighty-five percent of the population by 1901.

Close Window

Dan Tappan, New Mexico State University, Computing Research Laboratory

Toward Automated Scene Generation from Textual Descriptions

Natural language communicates descriptions of the world. Humans can decompose a complex visual scene into salient details, represent it with relatively few words, transmit it in written or verbal form, and then effortlessly reconstruct it with high fidelity. Very little information is actually stated, so humans rely heavily on commonsense knowledge and reasoning to fill in the gaps. Together, this explicit and implicit information helps the receiver build and manipulate a corresponding mental model of the scene.

Text understanding by computers is generally limited to superficial processing of grammar and vocabulary only. Consequently, most systems cannot benefit from important contextual cues to narrow interpretations, reduce ambiguity, and so on, and their performance suffers accordingly.

This prototype computational-linguistics system works toward bridging the gap between human and computer language processing. It employs concepts from linguistics, psychology, computer science, and related areas to convert textual descriptions into plausible visual interpretations.

Three major lexical-semantic issues are addressed:

Formal reasoning and inference methods are applied across these issues in conjunction with a source of relevant background knowledge. The final output is a simplistic graphical rendering of a virtual-reality environment that roughly corresponds to the textual description. The display is augmented with further information to account for alternative interpretations.
Close Window


Elisabeth Villalta, Frank Wijnen, Ted Sanders, Utrecht Institute of Linguistics

The Influence of Discourse Structure on the Processing of Quantifiers

Language users understand a discourse by constructing a coherent representation of the information in the discourse. One of the crucial ways to do so is by inferring coherence relations between segments (Sanders & Spooren 2000). In this paper, we investigate the hypothesis that principles of anaphora resolution fall out of the inference processes underlying establishment of discourse coherence relations (Hobbs 1979, Kehler 2002). We focus on one type of anaphora resolution: the resolution of the domain restriction of a quantifier. A self-paced reading study in Dutch provides first evidence that the resolution of the domain of a quantifier depends on the discourse-structure in which it appears.

Quantifiers (i.e., elements such as >many=, >most=, >three=, etc.) are anaphoric in nature given that they are necessarily interpreted with respect to a contextually determined set (Westerstahl 1984). The question from a sentence processing perspective is then: how does the parser determine the appropriate antecedent-set when more than one antecedent is available in the context? For example, in (1), 'three' may refer to a subset of the ten students or rather to a subset of all students.

  1. All students went to the conference. Ten went to the morning session. Three went to the afternoon session.
The resolution of the restriction of a quantifier has been argued to obey Forward Directionality (Hendriks & de Hoop 2001): the domain restriction of the quantifier is preferably interpreted as the last mentioned set. The psycholinguistic motivation for this principle is that it speeds up processing because the parser attempts to interpret the domain as the last activated set. Sentences with a quantifier obeying Forward Directionality should thus be read faster than sentences with a quantifier that does not obey Forward Directionality.

In this study, we examined whether the preference for Forward Directionality is in fact dependent on the type of discourse structure in which the quantifier appears. In a self-paced reading study in Dutch we measured reading times for sentences with a quantifier in two types of discourse-structures: short stories with a narrative structure (containing temporal connectives) versus the same stories minimally modified to a list structure (containing sentences with parallel structure, no temporal connectives). These stories were followed by a sentence with a quantifier that obeyed Forward Directionality (i.e. >three' in (1)) or by a sentence that did not obey Forward Directionality (i.e., if >three' were replaced with >twelve' in (1)). For an example of material, see the second page of the abstract.

Narrative structures consist of a temporal sequence of events and induce the expectation that the story will continue to be about the same set of entities. In the context of a narrative story, we thus expected that forward directional sentences would be read faster than non-forward directional sentences. For list structures, there is no reason for such an expectation. Hence, we expected there to be no difference between forward directional sentences and non-forward directional sentences. First results with 28 participants indeed support this hypothesis: for narrative stories, the difference in reading times between the forward directional condition (2204ms) and the non-forward directional condition (2353ms) was twice as large as in the case of list stories (Forward Directional Condition: 2185ms, non-Forward Directional Condition: 2244ms).

Upon completion of the study we expect to be able to show that the Forward Directionality Effect disappears in the context of a list structure and only remains a preference in the context of a narrative discourse structure. This would provide first experimental evidence supporting a coherence-driven approach to anaphora resolution in which the principles of anaphora resolution fall out of the inference processes underlying establishment of discourse coherence relations (Hobbs 1979, Kehler 2002).

Close Window


Douglas C. Walker, University of Calgary

French aspirate-h isn't aspirate, isn't <h> and isn't French

Despite its name, the phenomenon commonly referred to by the term 'aspirate-h' in Standard French (exemplified in such words as le héros 'the hero' and la housse 'the cover') does not involve aspiration. Nor is it restricted to items beginning with the letter <h> – a variety of forms manifest the same type of behaviour. Finally, both historically and synchronically, much of the evidence for this class of forms is not French in origin.

Traditionally, aspiration refers to a subset of words beginning with orthographic <h> which fail to undergo both elision and liaison and which require suppletive behaviour in certain determiners or adjectives. In this paper, I will (i) review the complexities involved in "aspiration behaviour", showing how it is much more heterogeneous and variable than standard accounts imply (and that it demonstrates the lack of internal cohesion in elision and liaison themselves) and (ii) argue that in addition to acting as disjunctive contexts or syllable islands (Tranel 1993: 810), aspirate-h words are subject to frequency and distance effects that will require larger scale quantitative studies before the phenomenon is fully understood.

Close Window

Eric Russel Webb, Western Michigan University

From [R ] to [R] via [z]: rethinking the history of /r/

The historical shift from apical [R ] to dorsal [R] is hardly remarkable in Romance and Germanic; this transition is noted in standard forms of French, German, Danish and Swedish, as well as in dialects of Dutch, English, Italian and Portuguese. The history of French provides a particular example of rhotic shift: [z] is attested as a rhotic variant during the seventeenth century, although this situation was short-lived and dorsal [R] eventually "won out." In this paper, I re-examine phonological change and seek to answer two fundamental questions posed by the French data. Firstly, why do some changes succeed when others are only temporary? Secondly, what are the larger, systemic ramifications of sound change?

The first portions of this paper comprise a presentation of the French data and a brief overview of the working theory of relational phonology (introduced in Russell Webb 2002), its historical antecedents and analytical methods. This approach states that phonological units can be defined as discriminable components of a larger system, where the articulatory and perceptual natures of sounds provide for emergent distinction and similarity within an organic—or natural language—whole. In a subsequent section, I provide relational, articulatory and perceptual definitions of the French [R ], [R] and [z]. A relationally defined phonological segment is an organically determined element whose singularity is derived from the relations it entails with other system members (in this case, the French continuant inventory) and whose structure emerges from these relations, rather than being universally available and locally determined. While such multiplex relations are not assumed to be the catalyst for sound change, a relational understanding of the sounds in question offers a compelling explanation as to why certain segments are susceptible to engage in paradigmatic operations of replacement or substitution and why others are resistant to such. In final part, I suggest that the systemic needs for discriminability and focalization are at the heart of both the temporary confusion of rhotics with [z] and the eventual exclusion of [z] as a viable alternative. Discussion of sound change from a relational perspective considers first and foremost the sound system within which it occurs, noting that any change in systemic relations—such as the composition of one system member—must respond to the greater needs for minimal distinction and maximal focalization (perception), as well as low-cost gestural patterns (articulation). Using the relational hypothesis as a background for a revisited explanation of apical-to-dorsal rhotic alternation, I demonstrate how a relational understanding of phonological segments enhances our understanding of sound change. This approach has the added benefit of providing a phonetically grounded evaluation of synchronic phenomena, such as allophony and free variation.

Close Window

Janet Wilson University of Texas, Arlington

The Present Tense in Kuche Narrative Discourse

Kuche, a Plateau language (Benue-Congo) of Northern Nigeria, has a verb form that in many respects matches Longacre’s (1990:109) characterization of a consecutive tense: it is relatively unmarked and it carries the primary storyline of a narrative. However, it is also widely used in various kinds of discourse, especially in foregrounded clauses. That is because its tense and aspect are interpreted according to the marking of the verb or verbs that precede it; it is variously interpreted as past, future, and even habitual, depending on the context. It is marked only for subject agreement and is morphologically identical to the Simple Present Tense. Its use is similar to what Kiparsky (1968:32-34) calls "conjunction reduction," which occurs in many early Indo-European languages: a series of verbs in a sentence may begin with one in aorist tense or in future tense, but the verb(s) following the conjunction is a present tense form. The difference here is that the Kuche form is used over long stretches of speech, often joined together by conjunctions, but occasionally not.

In this study, the clauses of two typical Kuche narratives are sorted into narrative clauses (as defined in Labov 1972), dialog/conversation, and non-narrative clauses (summaries, repetitions, flashbacks, states, etc.). Nearly 85% of the narrative clauses and 44% of the non-narrative clauses use the unmarked verb form, but they are invariably translated into English as past tense. Perfective aspect verbs are used in the first few narrative clauses, but only the unmarked verb form is used to narrate the rest. I argue that this use differs significantly from European "historical presents" in at least three respects: (1) the perfective form is almost never found in narrative clauses; instead, the unmarked form is used; (2) it is not interpreted as "past more vivid"; and (3) the same form may follow verbs of any tense-aspect marking and then be interpreted according to the preceding verb.

I conclude that Bolinger’s (1947) characterization of the Simple Present Tense as unmarked for time is applicable to Kuche as well as English, and may have even wider application. He says,

We might call the simple present tense the BASE TENSE, to which all other tenses are oriented but which is itself oriented to nothing, expressing merely the FACT OF PROCESS. The simple present . . . is ‘timeless’ not in the sense of ‘eternal’ but of ‘non-committed about time.’ Whenever, then, the speaker wishes to avoid the confinement of time implicit in other tenses, he uses the simple present. Bolinger (1947:436)

It is the semantic "non-commitment" to any specific time that allows the Kuche Simple Present Tense to absorb, if you will, the tense-aspect of preceding verbs.

  Return to HDLS-V Homepage