Grammatical Judgement Task for ASL

This test was developed within the framework of a research project that investigates the effects of age regarding the acquisition of American Sign Language (ASL) with regard to its grammatical processing (Boudreault, 1999; Boudreault & Mayberry, 2000).

Thirty deaf subjects, aged 18 to 84 with ASL as their primary language, participated in the study. The subjects were divided into three groups depending on their age of first contact with ASL. The groups were named (1) ‘native’, deaf people where one or both parents used ASL, (2) deaf people who learned ASL in the school, form age four to seven, and (3) late learners, who started to learn ASL between age eight and thirteen.

This is a receptive test. The subjects saw 168 ASL sentences and had to judge whether the sentences were grammatical or ungrammatical. The sentences used had an increased syntactic complexity. The subjects’ response accuracy and latency were measured.

All types of sentences consisted of fourteen grammatical and fourteen ungrammatical sentences. The 28 sentences for each task had a total number of morphemes ranging from 184 to 220. The subjects were tested on six types of sentences.

(1) Simple sentences: The sentences consisted of only uninflected signs, that is, only with plain verbs with no agreement. No grammatical facial expression or other kinds of expression were used. No agreement of loci, except for the pronouns in signing space, was used. The sentences were made ungrammatical by moving the verb to a different position in the sentence.

(2) Negative sentences: Only uninflected signs, except for the negative marker, were used. No agreement of loci, besides the possessive pronouns in signing space, was used. Two types of negative inflection were used: (1) the ASL sign NOT before the verb and (2) the negative non-manual facial morpheme. They were made ungrammatical by moving the sign NOT to a different location in the sentence. The negative non-manual facial expression appeared at the beginning of the sentence and continued during the sentence before the verb started.

(3) Directional verb sentences: Only uninflected signs with one verb that was inflected for person and number were included. The verbs used were directional, using the signing space to indicate person and number. Two types of directional verbs were used: (1) body-anchored verbs and (2) unanchored verbs. The body-anchored verbs do not take person or number inflection. The unanchored verbs take inflection for person and number. They need two persons for verb agreement and agreement among first, second, and third persons. The sentences were made ungrammatical by moving the verb phrase (i.e. verb and person/number inflection) to another phrase.

(4) Wh-sentences: The Wh-(question)-sentences consist of an uninflected sign and a Wh-marker. Half of the sentences had a verb that was inflected, and half of the sentences had plain verbs. There are two types of Wh-markers: (1) Wh non-manual facial morphemes (without any Wh-sign) and (2) the sign WHY and WHO at the end of the sentence. The sentences were made ungrammatical by moving the Wh-facial marker or sign to another phrase.

(5) Relative clause sentences: A relative clause (RC) sentence consists of two verbs. The verbs were either inflected or plain. There are two types of relative clause markers: (1) relative clause facial morphemes and (2) THAT/ITSELF sign markers. An RC subordinate is made with a facial expression only and without the RC sign. The RC markers (facial expression) were placed in the first part of the sentence, with the RC sign in the second part of the sentence. The sentences were made ungrammatical by switching the RC facial marker and its clause to the second part of the sentence, and by moving the RC sign to an earlier phrase.

(6) Classifier sentences: The classifier sentences consist of two clauses. The second clause used a verb of motion. Three types of classifiers were used: (1) CLASS-1 (animate and vehicles), (2) CLASS-2 (inanimate and object), and (3) SASS. The sentences were made ungrammatical by scrambling the spatial order of the classifiers.

The ASL stimuli vary in their degree of complexity. The chosen grammatical structures were based on ASL research and acquisition studies. The stimuli were developed by the deaf investigator of the study and another deaf native signer. A total of 168 sentences were presented. A pilot study was conducted. The ASL sentences were videotaped and shown to three ASL signers who judged if the sentences were grammatical or ungrammatical. The three judges agreed on the majority of the stimuli. The three judges disagreed whether particular sentences were ungrammatical. The stimuli were changed until all three ASL signers agreed that they were ungrammatical. Following the pilot, the stimuli were videotaped. Only signs with a high frequency were used to avoid confusion.

The videotaped stimuli were edited and subsequently recorded onto a CD-ROM and presented via a computer. A command pad (game pad) was attached to the computer during the testing. Two of the four buttons were used to indicate whether a sentence was grammatical or ungrammatical. The computer recorded the response accuracy and latency of the subjects. In doing so, the scoring was efficient, since the data was already in the computer.

American Sign Language was used during testing. The subjects were tested individually. The testing time varied between fourteen minutes and one hour. Prior to the test, an interview to obtain some background information on the subjects was conducted. After the interview, a practice session with eight ASL stimuli was done. The subjects were instructed to focus on syntactic structures of the sentences (facial expression, sign order, spatial arrangement) and to judge their correctness. The test was divided into four parts, with a three to five-minute break between each part. Each part consisted of 48 stimuli, randomly drawn from 168 stimuli. This test is not available since it only had been used in a research context so far.

The ASL Grammatical Judgement Task has also been used in further studies investigating ASL competence of bilingual deaf signers and how this impacts reading skills (Chamberlain & Mayberry, 2006) and how syntactic processing has an impact "that the onset of first language acquisition affects the ultimate outcome of syntactic knowledge for all subsequent language acquisition" (Boudreault & Mayberry, 2006, p. 608).

Among the strengths of this test are that (1) it has an easy, comprehensible testing format and (2) has an efficient analysis through its computer based format.

The ASL Grammatical Judgement Test has been adapted to  British Sign Language.

From: Tobias Haug: “Review of Sign Language Assessment Instruments”, an earlier version of that paper 2005.

For more information regarding this test, please contact  Patrick Boudreault at Gallaudet University, Washington DC.