Back Translation: A Method for the Analysis of Manual Languages

BACK TRANSLATION: A METHOD FOR THE ANALYSIS OF MANUAL LANGUAGES' Ryan D. Tweney and Harry W.Hoemann Studies of language functions in deaf persons have been hampered by the lack of an adequate formal grammar of American Sign Language (a lack which is undoubtedly based on the tendency to regard sign as "ungrammatical"). While Stokoe (1960) has described the gestural system used by American Sign Language (ASL) in terms of gestural phonemes or "cheremes"-minimal movements and gestural actions which will make a contrastive difference in meaning-few attempts have been made to describe formally the grammatical properties of ASL. McCall's (1965) attempt to write a formal grammar of ASL was hampered by two procedural shortcomings: (a) she attempted to write a grammar that would describe all utterances collected during a social gathering; thus, no selection of well-formed from badly-formed (or incomplete) utterances was made, complicating the task of description. (b) McCall's grammar applies an outmoded model of transformational grammar (Chomsky 1957). As important as that model has been historically, it suffers as a descriptive instrument, especially in comparison to later transformational grammars. For example, McCall uses the formal device of deriving one sentence from another via transformations while trying to specify "kernel" sentences in the language. In effect, this procedure allows derivation of any ASL sentence from a large number of other ASL utterances and does nothing to specify the underlying structure of the language. In such a system, "anything goes" as far as ASL word order is concerned. 1. This investigation was supported by Research Grant NS-09590-02 from the National Institutes of Health. Sign Language Studies The lack of an adequate formal grammar of ASL should not, however, prevent investigation of its relevant properties. Comparisons between the structure of ASL and English can be made without formal grammars, and can shed light not only on the properties of ASL as a linguistic system but also on its role as a communication medium. Moreover, contrastive study of ASL and English can have considerable practical as well as theoretical importance, since many educators have taken the relatively extreme position that communicative competence in ASL will interfere with deaf children's mastery of English (Lewis 1968). The present investigators adapted a method known as "back translation" to the study of ASL-a method described by Werner and Campbell (1970) as an adjunct to the translation of personality tests into a foreign language. Suppose that a psychologist is interested in translating a test into another language for which translators are available. It is well known that exact translation from one language to another is, in general, extremely difficult, if not impossible, since both linguistic and cultural differences can lead to many instances in which meaning is lost or distorted. Werner and Campbell approached this problem by using two (or more) translators in the following way. One translator would translate an English test item into the target language. The second would then take the translated item and translate back into English. The two English versions (original and back translation) were then compared. If both were semantically equivalent then the target language version was considered adequate. If they were not equivalent, consultation between translators and the test author was employed to uncover the source of translation difficulty, a new version of the item was written, and a second cycle of back-translation used to evaluate the new item. The advantage of the method is that it can generally reveal differences in meaning between the original and the target language translation, and is thus a powerful validation device. Since ASL has been characterized as a deficient language on the basis of a number of a priori and sometimes ethnocentric criteria (cf. Fusfeld 1958; Myklebust 1964; Tervoort 1961), there isobvious need for a reliable and empirical test of the adequacy of ASL as a communicative channel. Tweney & Hoemann Back translation is an appropriate procedure for this purpose since it can focus on the preservation of meaning when material is translated from one language into another. Additionally, back translation ought to be a revealing source of information on channel properties of ASL independently of message content, for whenever a translation embodies...