Showing papers on "Head (linguistics) published in 1994"

Subjects in Japanese and English

Abstract: Originally published in 1994, this volume shows that the structural relation 'government' holds not only between the verbal head and its object but also between the verbal head and its subject at least at the level of Logical Form in both Japanese and English. The book provides an analysis of complex predicate constructions in Japanese, discusses phrase structure in Japanese and English and develops a theory of binding.

Co-occurrence vectors from corpora vs. distance vectors from dictionaries

Abstract: A comparison was made of vectors derived by using ordinary co-occurrence statistics from large text corpora and of vectors derived by measuring the interword distances in dictionary definitions. The precision of word sense disambiguation by using co-occurrence vectors from the 1987 Wall Street Journal (20M total words) was higher than that by using distance vectors from the Collins English Dictionary (60K head words + 1.6M definition words). However, other experimental results suggest that distance vectors contain some different semantic information from co-occurrence vectors.

C-selection as feature-checking

Abstract: . Subcategorization (Chomsky 1965), or e-selection (Pesetsky's 1982 term), is a relation which allows a head to specify certain feature values on its complement. However, the limitations of subcategorization are identical to the limitations of head-movement; the relation is strictly between a head and the head(s) that it properly governs, and never between a head and a specifier or adjunct. The grammar can therefore be simplified by collapsing the two head-head relations, either by treating cases of abstract head movement (and concomitant feature-checking) as subcategorization or by treating cases of subcategorization as head movement. In this paper I outline a proposal of the latter sort.

Optimal contributions of head and eye positions to spatial accuracy in man tested by visually directed pointing.

Abstract: Encoding of visual target location in extrapersonal space requires convergence of at least three types of information: retinal signals, information about orbital eye positions, and the position of the head on the body. Since the position of gaze is the sum of the head position and the eye position, inaccuracy of spatial localization of the target may result from the sum of the corresponding three levels of errors: retina, ocular and head. In order to evaluate the possible errors evoked at each level, accuracy of target encoding was assessed through a motor response requiring subjects to point with the hand towards a target seen under foveal vision, eliminating the retinal source of error. Subjects had first to orient their head to one of three positions to the right (0, 40, 80°) and maintain this head position while orienting gaze and pointing to one of five target positions (0, 20, 40, 60, 80°). This resulted in 11 combinations of static head and eye positions, and corresponded to five different gaze eccentricities. The accuracy of target pointing was tested without vision of the moving hand. Six subjects were tested. No systematic bias in finger pointing was observed for eye positions ranging from 0 to 40° to the right or left within the orbit. However, the variability (as measured by a surface error) given by the scatter of hand pointing increased quadratically with eye eccentricity. A similar observation was made with the eye centreed and the head position ranging from 0 to 80°, although the surface error increased less steeply with eccentricity. Some interaction between eye and head eccentricity also contributed to the pointing error. These results suggest that pointing should be most accurate with a head displacement corresponding to 90% of the gaze eccentricity. These results explain the systematic hypometry of head orienting towards targets observed under natural conditions: thus the respective contribution of head and eye to gaze orientation might be determined in order to optimize accuracy of target encoding.

Method and apparatus for the determination of azimuth error

Abstract: Azimuth error between a write head (16) and a read head (34) is determined by passing a multi-track tape (2) past the write head (16) and subsequently the read head (34). The signals written by the write head (16) and read by the read head (34) are analysed to determine the timing error (38) between best-match time-delayed portions (20-28) of these signals. This timing error (38) is indicative of the azimuth error.

A Lexicalist Analysis of Gerundive Nominals in English

Abstract: Gerundive nominals in English are true categorial hybrids, combining a verbal complementation pattern with a characteristically nominal distribution and a clausal interpretation. Initial generative descriptions of gerundives attempt to resolve this conflict by assimilating these constructions to unambiguously nominal or clausal categories, or by assigning the discordant properties to different derivational levels. The alternative articulated here follows Schachter 1976 and Jackendoff 1977 in associating these conflicting properties with different projections in a phrase structure analysis. Gerundives are analyzed as nominal phrases headed by present participles, which, as verbal categories, maintain a predominantly verbal argument structure and interpretation. Following Pullum 1991, the clash between the category of a gerundive phrase and its lexical head is not simply stipulated, but rather attributed to the default projection of constitutive ‘head’ features, as proposed in extended phrase struc...

Syntactic-head-driven generation

Abstract: The previously proposed semantic -head-driven generation methods run into problems if none of the daughter constituents in the syntacto-semantic rule schemata of a grammar fits the definition of a semantic head given in [Shieber et al., 1990]. This is the case for the semantic analysis rules of certain constraint-based semantic representations, e.g. Underspecified Discourse Representation Structures (UDRSs) [Frank and Reyle, 1992].Since head-driven generation in general has its merits, we simply return to a syntactic definition of 'head' and demonstrate the feasibility of syntaclic-head-driven generation. In addition to its generality, a syntactic-head-driven algorithm provides a basis for a logically well-defined treatment of the movement of (syntactic) heads, for which only ad-hoc solutions existed, so far.

A theory of mind is in the head, not the heart

Abstract: menced these displays, including vocal "threats," when food was introduced into the test, and especially when the image of their hands came close to that of the food. (Menzel et al. 1985, p. 216) One account of this difference between chimpanzees and monkeys is that the latter simply do not realize that the hands they see in the mirror are their own (see Gallup 1991). C&S assert that mirror self-recognition is a narrow definition and that there must "be more than one criterion by which to measure the ways individuals in any species recognize themselves" (p. 174). The important issue, however, is not whether there are other ways that species can perceive themselves, but whether they can conceive of themselves at all. As C&S acknowledge, even the immune system shows a rudimentary form of self-recognition in terms of being able to identify and respond to foreign proteins. Likewise, many different species undoubtedly have a proprioceptive sense of self that makes it possible to take into account their own movement-induced experiences. But self-perception does not guarantee self-conception (Butterworth (1992). I have suggested that the reason monkeys fail to recognize themselves in mirrors is because they cannot conceive of themselves, and I have developed a model that identifies a variety of additional criteria that could be used to cross-validate the presence (or absence) of self-conception in other species (Gallup 1983; 1985; 1991). Of course C&S are free to continue to minimize mirror data, but the fact remains that I predicted their basic findings about the inability of monkeys to reason about mental states (either in themselves or other monkeys) many years ago (Gallup 1982). Not only does my model provide an account of their data, but it makes a number of other predictions (both between and within species) that have begun to receive support from the elegant work of Povinelli and his colleagues (e.g., Povinelli & deBlois 1992b; Povinelli et al. 1990; 1991; 1992a; 1992b).

Verb + Noun Function-Describing Compounds

Abstract: In this study we examine a class of exocentric nominal compounds (i.e. compounds with an unexpressed noun head) in Spanish, French, and Chinese. This class consists of nominal-compounds formed by a verb-plus a noun complement, usually though not necessarily a direct object, which combine to describe a function or characteristic of a new whole. In the three languages studied: here, compounds of this type tend to fall into two semantic groups: (1) utilitarian objects, such as 'paperweights' and 'armrests', which are perhaps best and most easily described by their functions; and (2) specialized professions, like 'drivers' and 'switchmen'; plus a subcategory of often pejorative, tongue-in-cheek descriptions of certain types of people, like 'wet blankets' and 'quack dentists'; and a smaller subcategory of other living things, i.e. plants and animals. The fact that languages in such diverse families as Indo-European and Sino-Tibetan-selectively, comparing, however, only SVO languages-exhibit the same type of compound, and use it to indicate extremely similar referents, suggests that certain objects in which a main function or peculiar characteristic stands out (e.g. an umbrella, whose one main function is to protect one from the rain) are more likely than words not of this type to be expressed in a verb + noun exocentric compound, in languages in which this compound type commonly occurs. This is also true of profession names, in which function is an outstanding element; as well as of certain types of people, who are identified chiefly by a particular characteristic (such as being a 'fight-picker'); and other living things possessing some salient feature. The study of this method of compounding-and compounding in general-may also have significance in the area of word order typology. The observation that the syntax of a language is reflected in its morphology is borne out through a comparison of this word type in languages not treated in this paper; for example, a similar compound type occurs in Burmese and Persian, both SOV languages, but in an inverted noun + verb format. Further examination of this compound type in various languages promises to make contributions to the study of language universals, word order typology, comparative morphology, and our understanding of the interrelationships among the three.

Que es un good code-switch? Testing the Functional Head Constraint within Noun Phrases

Abstract: The Functional Head Constraint (Rubin and Toribio, 1994) states that no code-switch is allowed between a functional head and its complement. This was tested in an experiment with Spanish/English code-switched sentences. Subjects in the experiments were native speakers of Spanish (N=21), who were proficient English speakers, and had learned English during adulthood. The experiment used the Matching Task described in Freedman and Forster (1985). In this procedure, two sentences are presented sequentially on a computer screen, and are kept on the screen for 2 milliseconds. Subjects press a "yes" button or a "no" button to indicate whether the sentences on the screen are identical. In this experiment, three conditions were used: condition 1, where the determiner and the noun were Spanish; condition 2, where the determiner was Spanish and the noun was English; and condition 3, where both the determiner and the noun were English. Grammaticality of the code-switch was established by comparing matching times on all SAME items for the two conditions. Here, condition 2 violates the functional head constraint; thus, matching times for this condition should be significantly slower than for conditions 1 and 3. The results obtained are interpreted under the Minimalist Program of Chomsky (1992).

Functional categories and Spec-head agreement

Abstract: In this work, we show that "doubling phenomena" in Romance reduce to Spec-head agreement relation within the phrase. Spec-head agreement is a licensing mechanism, part of UG, under which the doubled elements must be in a Spec-head configuration with each other. The discussion revolves around Romance languages primarily, although examples from other language families (eg Germanic) will be brought in. This study will concentrate on SPEC-head agreement which takes place within the maximal projection of functional categories, such as CP, AGRP-s, DP, NegP and AGRP-o. This presupposes the view proposed in recent GB literature that IP splits into further functional categories, such as subject and object agreement phrase and a Negation phrase. We also refer to the determiner as a functional category, heading its own projection, DP. The range of doubled structures we survey include: subject clitic doubling in some Northern Italian dialects, negative doubling in French and Portuguese Creoles, and object clitic doubling in River Plate Spanish. A doubly-filled Comp is attested in some French varieties as well as in some Germanic dialects. Agreement of the complementizer with the head NP of the relative clause in French ("que-qui rule") can also be reduced to Spec-head agreement. A solution to clitic doubled object NPs in terms of an AGRP-o allows us to abandon previous base generation analyses proposed for this construction. The clitic is considered not an argument but an affix-like head which can attach to a higher head, base-generated in Agr-o. The clitic, which shares phi-features with the object phrase, can double the latter because they stand in a SPEC-head configuration.