At present, researchers throughout Europe use a variety of transcription systems, and their representations of data differ accordingly. While it is inevitable that transcription systems will continue to vary, there is a clear need for a common database format. The aim of this chapter is to assess psycholinguistic evidence that would justify the inclusion of iconicity in such a database. That is, for each sign in the database, should there be some indication of the iconic motivation of the sign?
The chapter starts by attempting to define iconicity. A multitude of terms abounds, each often meaning something subtly different. Iconicity can refer to the extent to which a non-signer can guess a sign's meaning (transparency) or to the extent to which a relationship between a sign and its referent is apparent to a non-signer (translucency). It may also refer to the (historical) motivation of a sign form, whether or not that motivation is apparent to the perceiver. Clearly, using only one definition is not viable. Yet one must be aware of the different perspectives when assessing individual arguments.
The second section looks at how phonological theories have treated the issue of iconicity. Early work by William Stokoe established a set of formational elements that could be used to describe signs independently of their motivation (Stokoe, Casterline & Croneberg, 1965). This helped draw a distinction between a sign's form and its meaning. This distinction is one that many saw as essential to establishing sign languages as true language systems. This separation of form and meaning has been a principle of linguistic theory since Saussure and accords with psycholinguistic models of spoken language processing, where semantics does not feed directly into phonological form (see below). This approach dominates to the present day, with the work of Brentari, Sandler, Liddell, Perlmutter and others. More recently, however, a role for iconicity within sign phonology has been put forward by Armstrong, Stokoe and Wilcox (1995). Armstrong et al.'s arguments are based upon evolutionary arguments, and as yet there is little or no psycholinguistic evidence to support them. Some linguistic evidence for iconicity's role in sign phonology has been put forward by van der Hulst and colleagues (this volume), but the validity and necessity of these analyses have yet to be established.
A more salient role for iconicity lies in the acquisition of sign languages. No clear role in relation to deaf children has been established in the literature, but for hearing (second language) learners it has been shown that it is easier to learn and recall signs that have a clear motivational link to their referents. Presumably such a link provides a cue to retrieving the appropriate spoken language word from the learner's lexicon. The greater ease of learning signs with strong iconic imagery has been one of the bases for the use of signing with learning impaired children, although for this group, the advantage may be that children are actually learning enactive gestures rather than linguistic symbols.
The formational parameters of signs have been shown to influence short-term memory performance. Some researchers have argued for a sign loop in the short-term memory of native signers, analogous to the phonological loop in users of spoken languages (Wilson & Emmorey, 1997b). The evidence is persuasive but in need of replication. Only one study to date has used short-term memory paradigms to directly assess hypotheses generated from the proposals of Armstrong et al. (Dye & Shih, in preparation). The evidence obtained was mixed, yet the sample was not solely composed of native sign language users. In addition, there were considerable methodological complications in designing the study. At present, there is little evidence in the short-term memory literature to support a role for iconicity in sign language representation.
Finally, the evidence relating to iconicity from linguistics, language acquisition and cognitive psychology is summarised. It is concluded that the current literature establishes no clear role for iconicity in the phonological representation of signed languages. The salience of a sign's motivation is, however, apparent to those who learn sign languages as second languages. Only if a sign language database is to extend its remit to include data from second language learners does iconicity need to be considered. Even then, its role seems confined to memory and the acquisition of sign forms.
To an outside observer, sign languages seem to be an elaborate gesture system. Whilst most na•ve observers would claim that the signs of a sign language could easily be guessed, they are usually at a loss when exposed to a stream of fluent signing. However, when presented with isolated signs, na•ve observers can sometimes correctly guess the meaning of the sign. Whilst this is not true for all signs, it is nevertheless at odds with our experience of spoken languages. Given a word from an unknown spoken language, it is rarely possible to correctly guess the meaning of that word. Sign languages are therefore often referred to as being 'more iconic' than spoken languages. The aim of this section is to examine what we mean by 'iconic' when referring to individual signs of a signed language. We are confining ourselves here to single signs, and not to guessing the meanings of longer signed utterances. The early literature on sign language iconicity made heavy use of two terms: transparency and translucency. It should be recalled that these terms do not relate to measures of iconicity for individuals who know a sign language, but instead refer to tests of individuals who do not.
An individual sign's transparency refers to "the extent [to which] a sign's meaning can be understood from its form alone" (Klima & Bellugi, 1979, p. 22). Two ways in which the transparency of signs has been measured are (1) free-response guessing and (2) multiple-choice guessing. In both of these procedures, sign-na•ve subjects are presented with a set of sign language signs. They are required either to guess the correct meaning of each sign (free-response) or to select the correct meaning from a set of presented alternatives (multiple-choice). Klima and Bellugi (1979) report the results of both procedures using 10 sign-na•ve subjects and 90 ASL signs that were directly translatable into English. Using the free-response procedure, only 9 of the 90 signs produced any correct guesses. Using the multiple-choice procedure with the same signs and a new group of sign-na•ve subjects, performance was slightly less than chance for all subjects (an average of 18.2% given 5 alternatives to choose from). On only 12 out of the 90 signs did a majority of subjects choose the correct alternative.
Sign translucency refers to "the relation between a sign and its meaning" (Klima & Bellugi, 1979, p. 22) and how obvious this relation is to a na•ve observer. This is a more liberal measure of iconicity, in that the na•ve observer is provided with the sign as well as its translation into a language they do know (e.g. English). Their task is then to state the relationship between the sign and its referent. Using the same corpus of 90 ASL signs, Klima and Bellugi reported overall agreement amongst subjects on over half of the corpus. That is, although the meaning of many ASL signs could not be guessed, na•ve subjects could establish clear links between many signs and their referents. Interestingly, Klima & Bellugi note that the agreed relationships provided by their subjects often did not correspond to the historical motivations that gave signs their form. It is to historical motivation that we now turn.
As stated previously, it is clear that some signs are motivated by their referent. That is not to say that the referent determines the form of the sign. Was this the case, there would be greater correspondence between signs from different signed languages. A commonly cited example is that of various signs for "tree". Whilst the American, Danish and Chinese signs for "tree" are all iconically motivated, their actual forms within each of these signed languages is very different (Klima & Bellugi, 1979, pp. 21-22).
Using available records it is sometimes possible to trace the historical forms of current signs (see Frishberg, 1975; Kyle & Woll, 1985). In many cases this reveals a clear semantic motivation, i.e. the form of the sign is clearly derived from the sign's referent. One example given by Klima & Bellugi (1979, pp. 24-25) is the ASL sign GIRL. They state that: "According to historical sources the sign originally represented either the bonnet strings of hats worn by young girls or the curls that lay along their cheeks" (p. 24). In their study of sign translucency, many subjects reported that the sign took its form because of "the soft cheek of a girl" (p. 24). Presumably in the late 1970's few girls wore bonnets with chinstraps, and it was less common for girls to have ringlets of hair lying across their cheeks. The original motivation was not therefore available to the na•ve subjects, who produced a response more in line with their culture and experience (a point to which we return a little later). Such an example, and there are others, warns against inferring iconicity in signs that we observe; although a sign may have been motivated when it originated, that motivation may be quite different from the one that we infer now.
Signs change over time. The forms of signs in current use in BSL are not exactly the same as in 19th century Britain, even if we exclude neologisms such as WASHING-MACHINE. One possible cause of historical change is a move towards regularisation of newly introduced iconic signs (Bellugi & Klima, 1976). Newly developed signs may clearly resemble their referents iconically, yet in so doing violate constraints within the language on what form signs may take. As an example, Bellugi and Klima cite the ASL sign VIDEO. This can be produced in two ways. Whilst both versions are two-handed signs with G-handshapes positioned palm-down, fingertips away from the body at waist height, they differ in their movement. In one variant, both hands are moved in small clockwise circles in the horizontal plane; in the other, the left hand is moved in a clockwise direction and the right hand in an anti-clockwise direction. The first variant (and earlier form) more closely mirrors the actual movement of the reels on a video tape recorder. The latter form more closely conforms to ASL rules on the plane symmetry of two-handed signs, although in doing so loses some of its iconicity. It is possible that the degree of motivation in a language may be under-estimated. It is also important to note that the phonological rules that operate upon the form of signs within a language take precedence over historical motivation. The precedence of syntactic rules in relation to motivation is more complex.
In an example such as the noun TELEPHONE in BSL, the handshape represents a telephone handset, located at the ear. There is a derived verb TO-TELEPHONE, in which the hand moves from a location assigned to the subject to a location assigned to the object. Thus in I-TELEPHONE-HIM, the hand moves from the ear to a location assigned to the 3rd person. In SHE-TELEPHONES-HIM, the hand moves from one 3rd person location to another 3rd person location. In TELEPHONE-EACH-OTHER, both hands move alternately. Clearly, the movement of the hands does not represent any actual movement of a telephone handset.
On the other hand, some signs appear to retain features of their original motivation in relation to grammatical processes. Although size of a referent is indicated in many BSL signs by changing the size or extent of movement, signs in which a part stands for the whole referent, such as CAR (in which an imaginary steering wheel is turned) or PRIEST (in which a clerical collar is outlined), cannot be modified in this way. If the movement in PRIEST is enlarged, it means LARGE-COLLAR, not LARGE-PRIEST; while changes in the size of movement of CAR result in the meaning BIG-STEERING-WHEEL.
Another issue to be considered when assessing the iconicity of a sign language or its signs is the cultural bias that an observer brings. What is iconic to an English college student may not be so to a Polynesian fisherman (and vice versa)! When we attribute a degree of iconicity to a sign or signs, we are bringing to bear our own culture and experience. Another observer may make different attributions, or the iconicity may not be apparent to the very producer of the sign being assessed. Here are two examples, one relating to age (following Markowicz, 1977) and the other to culture.
The BSL sign MILK appears iconic to most adults, clearly resembling the action of milking a cow by hand. However, to a young urban child learning that sign, no motivation is likely to be apparent. The very fact that the sign has not changed over time to reflect changes in dairy practices, suggests that motivation is not an active process for a child learner.
A cultural example would be a BSL sign for FISHERMAN. While this sign may be transparent to anyone with experience of using a fishing rod, it would not be so to someone whose experience of fishing is based upon spearing or netting fish. Indeed, for such individuals, the sign would not even be translucent (unless it referred to an extremely unsuccessful fisherman). Such cross-cultural differences in sign translucency have been demonstrated within Europe (Pizzuto & Volterra, 1998).
When referring to iconicity we can mean different things. Iconicity may refer to the ease with which a sign's meaning can be guessed (transparency), the extent with which it can be associated with its referent (translucency) or the extent to which a sign was actually historically motivated. However we choose to interpret the term, there are at least two important things to keep in mind. Firstly, the rules of well-formedness within a sign language are more important than the iconic value of signs. Secondly, the iconicity of a sign is an attribution made by the observer. Such attributions are influenced by the experiences of the observer, which are determined by many factors, including age and culture. These factors are important to bear in mind in relation to the next sections, looking at iconicity within the framework of phonology, language acquisition and mental representation.
Phonology is the study of how elements of language (i.e. words and signs) are constituted. It seeks to formulate underlying representations of forms, and to specify rules about how these representations are manipulated in order to create the phonetic output that is produced by the signer and received by the perceiver. Ideally, these representations and rules should be parsimonious and capable of accounting for all observed data.
A starting point is the dissociation of linguistic and non-linguistic gestures. Linguistic gestures are componential, in that they are made up of sub-units combined in rule-governed ways. Non-linguistic gestures are holistic, and cannot be easily broken down into component parts. Consider two examples, one linguistic and the other non-linguistic, both of which represent an 'egg'. The first is the BSL sign EGG. Whilst this sign can be considered iconic within British culture (it can be recognised as representing the use of a spoon to crack the shell of a boiled egg), it is made up of two handshapes and a movement seen in other BSL signs with very different meanings. In addition, these elements (handshapes, movement) are combined in a lawful way (i.e. the combination of handshapes and movement is acceptable within the rules of BSL). A non-linguistic representation of 'egg' is also possible. It may be a pantomimic representation, acting out the placing of a boiled egg in an egg holder, cracking the top of the egg, slicing the top off with a knife and then eating it. Whilst this pantomimic representation is componential in that it is made up of a series of actions, those actions are closely tied to the actions required to perform the activities and are holistic in nature. Another important distinction between the lexicalised sign EGG and a pantomimic representation of egg relates to their connection with their referents. The sign EGG refers to any egg, and can be used whether the egg is raw, hard-boiled, scrambled or fried. The pantomimic representation specifically describes a hard-boiled egg, and a different pantomime would be required to make it clear that a fresh egg was being referred to. Such a distinction is made in BSL by the addition of another sign, so for example, 'boiled egg' would be represented by the addition of the signs BOILED or HARD.
A sign's form is not directly related to what the sign means. A sign is therefore free to vary in form, with its meaning agreed by convention. Deuchar (1991) argues that linguists' traditional focus on arbitrariness in the symbol-referent relationship is misleading, and that the important characteristic about a sign's relationship with its referent is conventionality: an (implicit) agreement amongst a community of language users on the meaning of a particular symbol.
Duality of patterning refers to the existence of a double layer of structure relating meaning and form. In spoken languages, for example, words are composed of phonemes (abstract sound units) which in themselves have no meaning. Only when they are combined in a particular order does a meaningful word occur. For example, the phoneme /s/ has no meaning in itself. It can be combined with /¾/ and /p/ to form 'asp', 'pass', or 'sap'.
A central question for the study of signed languages is whether they exhibit this duality of patterning, especially in the presence of iconicity. If duality of patterning can be shown to exist, then it is clear that there is no linguistic argument for the inclusion of iconicity in a phonological transcription database. If a link between form and meaning can be established, however, then the iconic value of a sign cannot be ignored so readily.
William Stokoe and his colleagues (Stokoe, Casterline & Croneberg, 1965) were the first to formulate a set of sub-lexical units for the signs of a signed language, namely ASL. He described signs as consisting of three main parameters (called 'cheremes'): dez (or handshape), tab (or location) and sig (or movement). Any sign could be formed from a simultaneous bundle of these parameters. For example, the BSL sign EGG could be described as a hooked index finger handshape (dez) that twice taps (sig) a fist handshape (tab). A finite set of such parameters was specified. Further research (Friedman 1978) described rules that determined how these parameters were combined to form all known and possible signs of ASL. Such analyses of ASL were used to establish that duality of patterning existed within ASL, and therefore that ASL was not just elaborate pantomime but a linguistic system worthy of scientific analysis.
Duality of patterning is also a component of modern approaches to sign language phonology. The Stokoe model of key parameters has been discarded in terms of more complex feature geometries (see Brentari, 1999; Liddell and Johnson, 1989; Perlmutter, 1992; Sandler, 1989). Modern approaches view signs as similar in structure to the words of spoken languages. In both cases, they are seen as combinations (both simultaneous and sequential) of linguistic elements. One of the first such approaches was that of Liddell and Johnson (1984), who argued that signs were made up of strings of movement and hold segments. Each segment consisted of a set of features that specified in detail the articulation of the sign. Movement segments were composed of features which detailed the hand(s)' movement through space as well as internal movement such as finger wiggling; hold segments contained features which detailed the articulatory arrangement of the hand(s). As in all formal phonology, the actual features are held to bear no relationship to the meaning of the sign produced; the fact that the little finger is extended (i.e. the sign contains the feature [+little finger]) does not relate to any property of the sign's referent itself. Whether or not modern theorists agree on the need to postulate segments (see Edmondson, 1986; Wilbur, 1990; Wilbur & Allen, 1991; Wilbur & Peterson, 1997), they all specify a set of features that are used to specify the form of signs within a signed language.
Recent work by linguists in the Netherlands (this volume) has questioned whether this featural approach is adequate to describe the phonology of signed languages. One way in which the features of a language are empirically established is via the identification of minimal pairs. A minimal pair is a set of two signs with different meanings, but which differ from each other in only one characteristic. Consider the example of a minimal BSL pair: (YOU'RE)-RIGHT and YOUR. These two signs have different meanings, but differ in only one respect: in the sign (YOU'RE)-RIGHT the thumb is extended. Within the Stokoe framework, this would indicate that a fist handshape is one dez of BSL while a fist handshape with the thumb extended is another. In modern featural theories [thumb] would be described as a distinctive feature, with (YOU'RE)-RIGHT taking the value [+thumb] and YOUR taking the value [-thumb].
One problem with this approach is that it leads to a very large number of distinctive features. The number for most signed languages is far in excess of that for spoken languages, and represents a problem for the featural approach to sign language phonology. Van der Hulst and colleagues argue that it is possible to utilise a much-reduced set of features if one takes into account the iconicity of signs. Within this framework, iconicity is seen as a factor that modifies what would be the 'normal' feature-determined form of a sign. For example, the Dutch Sign Language (SLN) sign PHARMACIST is articulated on the left side of the chest. Van der Hulst et al argue that this sign's features would predict its articulation on the sternum, but its iconic motivation, representing something being inserted in the breast pocket, leads to it being offset in its final articulated form. Van der Hulst et al argue that including iconic information in phonological representation means that far fewer features need to be postulated, as iconicity will account for signs which deviate from the predicted form. If this is so, then any transcription database must include an account of a sign's iconic motivation, in order for the discrepancy between phonological (featural) and phonetic (articulated) form to be accounted for.
Another recent theory of sign language phonology (semantic phonology) has also proposed a link between sign form and meaning. Stokoe (1991) and Armstrong, Stokoe and Wilcox (1995) have argued against any form-meaning distinction. Their argument is that signs are representations of entities performing actions. In their model, the handshapes, locations and movements that characterise a sign's form are all motivated by what the sign represents.
Consider again the BSL sign EGG. Within semantic phonology this represents one object acting upon another object. The active handshape takes the form it does, because the instrument of action is an implement with a long thin handle and curved extension (a spoon), and the location (non-active handshape) is a fist, because it represents an object with similar dimensions (an egg). The form of the sign is therefore directly motivated by its referent.
Such an analysis is reminiscent of Wallin's (1996) analysis of polymorphemic signs, where the sign is analysed as possessing a figure, motion and ground. The figure is the main agent or instrument in the representation, the motion is the action performed by the agent/instrument, and the ground is the location of the scene or the object of the action. In addition, handshapes are selected on the basis of their physical relationship to their referents. Thus a piece of paper would be represented with a flat handshape, with fingers extended and together, since both this handshape and a sheet of paper have large x and y dimensions, with little or no z-dimension. Although Wallin's analysis applies only to polymorphemic signs (and is thus compatible with duality of patterning), semantic phonology is proposed to apply to all signs (including monomorphemic ones).
At present, semantic phonology is not a fully developed phonological theory. It is not clear what the elements of a semantic phonology would be, and no attempt has been made to specify them. In addition, it has not been made clear how such a theory accounts for the form of a) signs with concrete referents but no apparent iconic motivation (such as BATTERY); and b) signs representing abstract referents that are clearly non-iconic (e.g. SISTER).
Whilst phonologists may argue over the relevance of iconicity to the representation of signs, psychological studies of sign language have addressed the topic in an experimental manner. One area of psychological inquiry that has attempted to address the role of iconicity is sign language acquisition. Bearing in mind statements in the introduction to this chapter on the salience of iconicity to those with different experiences, this section is divided into native acquisition by deaf children and acquisition by second language learners.
Orlansky and Bonvillian (1984) examined the vocabularies of ASL L1 learners. They observed the sign production of 13 hearing children of deaf parents, who were acquiring ASL as a first language. Items were classified as iconic, transparent or arbitrary (note that their use of these terms is different from that of Klima and Bellugi, discussed earlier). Iconic items were signs that looked like their referents; transparent productions represented a part or component of their referents (these would more usually be called hyponyms); and arbitrary productions bore no obvious relationship to their referents. Interestingly, Orlansky and Bonvillian found that signs produced in the first year were evenly distributed across all three categories. That is, L1 learners did not produce iconic signs more frequently than non-iconic signs. This suggests that for L1 learners signs are not tied to real-world referents; rather, they serve a symbolic, linguistic function (Marschark, 1993). This finding was supported by further studies by the same research group (Bonvillian & Patterson, 1993; Folven & Bonvillian, 1991). Although these studies are concerned with hearing L1 learners of ASL, there is no reason to expect a different pattern of acquisition by deaf children.
Much research has focused on the learning of signed languages by hearing adults, motivated by interest in the use of signs as an augmentative/alternative communication system for learning impaired children and adults. Hearing sign-na•ve subjects of normal intellect can identify the iconicity of many signs (Beykirch, Holcomb & Harrington, 19??; Granlund, Stršm & Olsson, 1989; Griffith, Robinson & Panagos, 1981; Hoemann, 1975), and can state the relationship between sign and referent (Griffith, Robinson & Panagos, 1981), although some signs are labeled incorrectly but with a consistent response across subjects (Hoemann, 1975). In addition, although both transparent and non-transparent signs can be retained over short and long periods of time, non-transparent signs are more likely to be forgotten when subjects are tested after a two week delay (Lieberth & Gamble, 1991). Sign translucency (in Klima & Bellugi's sense has also been demonstrated to be an aid to sign learnability by adults (Luftig & Lloyd, 1981; Mandel, 1977).
Another field of psychological inquiry that has examined the role of iconicity is memory research. Research in the 1960s into memory for spoken language established that short-term memory encodes linguistic input in terms of its phonetic form. Furthermore, longer-term representations tend to be semantic in nature, with meaning preserved but the exact phonetic form of the input lost. Working memory for language, the linguistic representations that are available to conscious awareness, has also been shown to be phonetic in nature. In the 1980's and 90's, researchers set out to determine the phonetic representation of signed languages in short-term memory.
A key experimental finding used to establish the phonetic nature of short-term memory coding, is the phonetic similarity effect. When a list of phonetically similar (e.g. rhyming) items is presented for immediate ordered recall, fewer items are recalled as compared with lists of phonetically dissimilar items (Baddeley, 1966). It is argued that phonetically similar items lead to phonetically similar representations, which are more easily confused at recall.
A seminal paper by Poizner, Bellugi and Tweney (1981) examined the effects of intra-list similarity on recall for native signers of ASL. Three types of intra-list similarity were manipulated: phonetic, semantic and iconic. Phonetically similar lists were composed of signs with similar handshapes; semantically similar lists consisted of signs from the same semantic class (e.g. fruit and vegetables); iconically similar lists consisted of signs that were high in iconicity, as rated by independent native ASL signers. Although semantic and iconic similarity did not lead to poorer recall, phonetic similarity did. This and other studies (Hamilton & Holzman, 1989; Klima & Bellugi, 1979; Kyle, 1986; Wilson & Emmorey, 1997a,b, 1998a,b) have led researchers to propose that signers represent signs in STM in terms of their handshapes, locations and movements. However, Poizner et al. (1981) did not actually manipulate iconic similarity. Rather, one list was high in iconic value, and the other low. Their data do not support the claim that representations have no iconic component. Rather, they establish that iconicity does not serve as an aid to recall in an STM recall task.
It is not at present clear what an iconically similar list would look like. How would one go about measuring the iconic similarity of two signs? Recent work by Dye and Shih (in preparation) has attempted to do this, utilising the arguments of semantic phonology. Lists were created that were motivationally similar or dissimilar, with motivational similarity defined on the basis of handshape. Signs in a motivationally similar list share a handshape because the signs in the list represent actions upon similar types of objects: the handshapes may be seen as size and shape specifiers (for example: SUITCASE, CAR, BRUSH, ICE-CREAM). For native BSL signers no effect of motivational similarity was found, although no effect of phonetic similarity (pure handshape similarity regardless of motivation) was found either. At present the literature is therefore inconclusive, although there is strong evidence that native signers use a (sign) phonetic code under some circumstances.
Armstrong, D. F., Stokoe, W. C. & Wilcox, S. E. (1995). Gesture and the nature of language. Cambridge University Press.
Baddeley, A. D. (1966). Short-term memory for word sequences as a function of acoustic, semantic, and formal similarity. Quarterly Journal of Experimental Psychology, 18, 362-365.
Bellugi, U. and Klima, E. S. (1976). Two faces of Sign: Iconic and abstract. In S. Harnad, D. Horst & J. Lancaster (Eds.). Origins and evolution of language and speech. New York: Annals of the New York Academy of Sciences.
Beykirch, H. L., Holcomb, T. A. & Harrington, J. F. (19??). Iconicity and sign vocabulary acquisition. American Annals of the Deaf, 135 (4), 306-311.
Bonvillian, J. D. & Patterson, F. G. P. (1993). Early sign language acquisition in children and gorillas: Vocabulary content and sign iconicity. First Language, 13, 315-338.
Brentari, D. (1999) A prosodic model of sign language phonology. Cambridge MA: MIT Press.
Deuchar, M. (1991). Are the signs of language arbitrary? In H. Barlow, C. Blakemore & M. Weston-Smith (Eds.) Images and understanding: thoughts about images, ideas about understanding (pp. 168-179). Cambridge: Cambridge University Press.
Dye, M. W. G. & Shih, S. (in preparation). Formational and motivational codes in the short-term memory of Deaf signers of British Sign Language.
Folven, R. J. & Bonvillian, J. D. (1991). The transition from nonreferential to referential language in children acquiring American Sign Language. Developmental Psychology, 27, 806-816.
Griffith, P. L., Robinson, J. H. & Panagos, J. M. (1981). Perception of iconicity in American Sign Language by hearing and deaf subjects. Journal of Speech and Hearing Disorders, 46, 388-397.
Hamilton, H. & Holzman, T. G. (1989). Linguistic encoding in short-term memory as a function of stimulus type. Memory and Cognition, 17 (5), 541-550.
Hoemann, H. W. (1975). The transparency of meaning of sign language gestures. Sign Language Studies, 7, 151-161.
Klima, E. S. & Bellugi, U. (1979). The signs of language. Cambridge, MA: Harvard University Press.
Kyle, J. G. (1986). Sign processes in deaf people in working memory. (MRC Report No. G 83-25637-N). University of Bristol: England. School of Education Research Unit.
Kyle, J.G. & Woll, B. (1985) Sign Language: the study of deaf people and their language. Cambridge University Press.
Liddell, S. K. & Johnson, R. E. (1989). American Sign Language: The phonological base. Sign Language Studies, 64, 195-277.
Lieberth, A. K. & Gamble, M. E. B. (1991). The role of iconicity in sign language learning by hearing adults. Journal of Communication Disorders, 24, 89-99.
Luftig, R. L. & Lloyd, L. L. (1981). Manual sign translucency and referential concreteness in the learning of signs. Sign Language Studies, 30, 49-60.
Mandel, M. A. (1977). Iconicity of signs and their learnability by non-signers. In W.C. Stokoe (Ed.) Proceedings of the First National Symposium on Sign Language Research and Teaching (pp. 259-266), Chicago, Illinois.
Markowicz, H. (1977). American Sign Language: Fact and fancy. Washington, DC: Gallaudet College Press.
Marschark, M. (1993). Psychological development of deaf children. Oxford, UK: Oxford University Press.
Orlansky, M. D. and Bonvillian, J. D. (1984). The role of iconicity in early sign language acquisition. Journal of Speech and Hearing Disorders, 49, 287-292.
Perlmutter, (1992). Sonority and syllable structure in American Sign Language. Linguistic Inquiry, 23 (3), 407-442.
Pizzuto, E. & Volterra, V. (1998).
Poizner, H., Bellugi, U. & Tweney, R. D. (1981). Processing of formational, semantic and iconic information in American Sign Language. Journal of Experimental Psychology: Human Perception and Performance, 7 (5), 1146-1159.
Sandler, W. (1989). Phonological representation of the sign: Linearity and nonlinearity in American Sign Language. Dodrecht, Germany: Foris Publications.
Stokoe, W. C. (1991). Semantic phonology. Sign Language Studies, 71, 107-114.
Stokoe, W. C., Casterline, D. & Croneberg, C. (1965). A dictionary of American Sign Language on linguistic principles. Washington, DC: Gallaudet College Press.
Wallin, L. (1996). Polysynthetic signs in Swedish Sign Language. Stockholm, Sweden: Department of Linguistics, University of Stockholm.
Wilbur, R. B. (1990). Why syllables? What the notion means for ASL research. In S. D. Fischer & P. Siple (Eds.) Theoretical issues in sign language research, Vol. 1: Linguistics. London, UK: University of Chicago Press.
Wilbur, R. D. & Petersen, L. (1997). Backwards signing and ASL syllable structure. Language and Speech, 40 (1), 63-90.
Wilbur, R. B. & Allen, G. D. (1991). Perceptual evidence against internal structure in American Sign Language syllables. Language and Speech, 34 (1), 27-46.
Wilson, M. & Emmorey, K. (1997a). Working memory for sign language: A window into the architecture of the working memory system. Journal of Deaf Studies and Deaf Education, 2 (3), 121-130.
Wilson, M. & Emmorey, K. (1997b). A visuospatial "phonological loop" in working memory: Evidence from American Sign Language. Memory and Cognition, 25 (3), 313-320.
Wilson, M. & Emmorey, K. (1998a). A word length effect for sign language: Further evidence for the role of language in structuring working memory. Memory and Cognition, 26 (3), 584-590.
Wilson, M. & Emmorey, K. (1998b, November). Modality matters: Spatial coding in working memory for signs. Paper presented to Theoretical Issues in Sign Language Research 6, Washington DC.
Posted: 24.01.2000