Perception & Psychophysics 1983,33 (3),215-231 Perception of movement in American Sign Language: Effects oflinguistic structure and linguistic experience HOWARD POIZNER TheSalkInstitute/or Biological Studies, La Jolla, California In order to reveal the psychological representation of movement from American Sign Lan­ guage (ASL), deaf native signers and hearing subjects unfamiliar with sign were asked to make triadic comparisons of movements that had been isolated from lexical and from grammatically inflected signs, An analysis of the similarity judgments revealed a small Ilet of physically specifiable dimensions that accounted for most of the variance. The dimensions underlying the perception of lexical movement were in general different from tholle underlying inflectional movement, for both groups of subjects. Most strikingly, deaf and hearing subjects significantly differed in their patterns of dimensional salience for movements, both at the lexical and at the inflectional levels. Linguistically relevant dimensions were of increased salience to native signers. The difference in perception of linguistic movement by native signers and by naive ob­ servers demonstrates that modification of natural perceptual categories after language acquisi­ tion is not bound to a particular transmission modality, but rather can be a more general consequence of acquiring a formal linguistic system. American Sign Language (ASL) is the visual­ and as a system with biological bases. In order to gestural language used by deaf communities in the bring these issues into focus, a brief description of United States. Since the language has developed com­ ASL will be presented first. pletely outside the auditory modality, its study can ASL is a language passed from one generation of provide clues both to the nature of language and to deaf people to the next as a native language. It is a those psychological processes upon which the com­ fully autonomous language with complex organiza­ prehension and production of language rest. The tional properties not derived from spoken languages. present article is concerned with how deafsigners and ASL is a highly inflective language much more like hearing subjects unfamiliar with sign perceive move­ Hebrew or Latin than like English (Bellugi, 1980; ments of the hands and arms that function in very Klima" Bellugi, 1979; Lane &: Grosjean, 1980; Siple, specific ways in the linguistic system of ASL. By 1978; Wilbur, 1979; Bellugi " Klima, Note 1). Like comparing perception of linguistic movement in sub­ all spoken languages, ASL shows duality of pattern­ jects who do and who do not know the language, we ing: basic lexical units are formed from a limited set can begin to uncover the ways in which linguistic ofcombining elements that are themselves essentially categorizations are based on the perceptual con­ without meaning, and meaningful units are com­ straints of the observer, and the ways in which such bined in the signed sentence by grammatical rules. categorizations are more arbitrarily determined. In However, the mechanisms by which this linguistic this way we can begin to address the correspondence structure is conveyed strongly bear the mark of the between language analyzed both as a formal system modality in which the language developed. With respect to the first level of patterning, signs from ASL are composed of three major formational This work was supported in part by National Institutes of parameters: configuration of the hands, movement Health Grant HD13249 and by National Science Foundation of the hands and arms, and location of the hands Grant BNS79-16423 to the SalkInstitute for Biolopcal Studies. I relative to the body (Stokoe, Casterline, &: Croneberg, thank Ursula Bellup for her valuable help throughout this project, 196'). Each parameter comprises an inventory of and Don Newkirk and Edward Klima for helpful discussions of the discrete representatives that combine concurrently paper. Malinda Williams, Crail Will, and Mike Herrera assisted with the experiment. I thank Emilio Bizzi, John Hollerbach, and rather than linearly. These representatives function Mike Atkenson for use of and assistance with the Selspot separately, however, to contrast minimally different mevement-monltoring system, which diptized the hand move­ signs, much as the phonemes of spoken language ments presented in Fiaure 3. Bell Labs kindly supplied the minimally contrast words. In the sign HOMEI, for SINDSCAL computer program, Portions of the experiment on lexical movement were reported in Sclencr, 1981, 212, 691-693. example, an O-shaped hand makes two contacts on Requests for reprints should be sent to Howard Poizner, The Salk the cheek. The sign HOME differs in form from the Institute, P.O. Box 85800, San Diego, California 92138. sign PEACH only in movement: in PEACH, the 0 215 Copyright 1983 Psychonomic Society, Inc. 216 POIZNER hand makes a circling movement on the cheek. Sim­ Bellugi, 1979, and Supalla & Newport, 1978, for a ilarly, other pairs of signs are minimally differen­ more complete description of these and other mor­ tiated by representatives of other formational param­ phological processes in ASL). Importantly, these eters. Furthermore, the formational parameters of modifications of signs are transmitted by superim­ signs not only describe linguistic structure but are posed changes in the movement of signs. also clearly important to the way signers process Fundamentally, the present study asks three ques­ signs(Bellugi, Klima, & Siple, 1975; Newkirk, Klima, tions about the perception of movement in ASL. Pedersen, & Bellugi, 1980; Poizner, Bellugi, & First, if the various movements at both the lexical Tweney, 1981). and the inflectional levels are perceived in terms of a The concurrent packaging of structural informa­ limited set of underlying dimensions, how might ob­ tion in ASL is also apparent at the morphological tained perceptual dimensions, for deaf and for hear­ level. With respect to the second level of pattern­ ing subjects, relate to the linguistic dimensions that ing in ASL, basic signs themselves undergo inflec­ have been proposed to account for relationships in tional processes that modify their meanings. These the linguistic system? processes impose regular changes in form across Second, do the dimensions that may underlie per­ syntactic classes of signs, and serveas a major vehicle ception of lexical movement differ from those that in ASL for expressingsemantic distinctions. Further­ may underlie perception of inflectional movement? more, inflections in ASL, as in speech, are not op­ If some perceived dimensions at the two linguistic tional but, rather, are required in sentential con­ levels differ, then data from language processing texts of the language (Klima & Bellugi, 1979). The would support the linguistic analysis that suggests distinctions between lexical units and inflectional that formational material at the two linguistic levels processes acting on them have also been shown to may in part be different. This situation would dif­ have psychological as well as linguistic validity fer from that of most spoken languages, in which the (Poizner, Newkirk, Bellugi & Klima, 1981). How­ same kinds of speech sounds are used to form both ever, unlike English, which typically inflects words basic words and their grammatical inflections, and by adding segments to the end of a word, such as the might point to ways in which the transmission past tensemarker [fJd] in needed, or the plural marker modality might shape the form that languages take. [s] in books, ASL inflects signs by modulating the Finally, and most importantly, does the psycho­ movement and spatial contouring of the sign. Bellugi logical representation of movement in ASL differ and Klima (Note 1)argued that the appropriate anal­ between native signers and naive observers? If so, ysis of morphological structure in ASL is in terms of then the linguistic experience of the deaf must have, multiple tiers: an underlying lexical root, and in­ in part, determined their perception of this forma­ flectional and derivational tiers overlaid concurrently tional parameter. Related experiments on the percep­ with the root. They pointed out that uninflected lexi­ tion of speech indicate that either natural nonlin­ cal stems are embedded in simultaneously overlaid earities of the auditory system or particular phono­ changes in movement and spatial contouring: a sign logical experience can determine a speaker's percep­ and its inflectional marker co-occur in time. Further­ tion of phonemes. As an example of the former, more, it appears that some of the building blocks and natural nonlinearities of the auditory system for dis­ combinations of building blocks used to construct criminating differences in the temporal order of lexical items and inflectional processes may differ. events have been linked to the perception of voic­ For example, a certain inflectional process embeds ing contrasts in speech (Aslin & Pisoni, 1980; sign stems in iterations along a circle of a horizon­ Jusczyk, Pisoni, Walley, & Murray, 1980; Pisoni, tal plane; another embeds sign stems in uneven re­ 1977). The following pattern of results serves as an duplicated elliptical movement. These (and other) example of the phonological determination of lan­ global movement contours do not occur as move­ guage perception. Human infants discriminate ments of basic uninflected signs(Bellugi, 1980; Klima acoustic differences that cue the distinction between & Bellugi, 1979). the phonemes Irl