A Language Of The Hands

AThat the signs used by deaf communities resemble the things they refer to was, for a long time, the main reason for denying them the status of language. If a sign for a bird traces a beak opening and closing, the reasoning ran, then signing is drawing rather than speaking, and a system of pictures cannot express what cannot be pictured. The premise was less secure than the conclusion. Resemblance in these languages is real, but it is selective and conventional: one language may represent a cat by its whiskers and another by the stroking of its fur, and a signer who has never met the second language will not reliably guess what the sign means. The old prejudice has in any case been replaced by a friendlier error, namely that resemblance makes such languages easy to acquire. Both mistakes share an assumption worth examining, that a form which can be explained by a picture is therefore produced and understood as a picture. Most research of the past two decades has been directed, one way or another, at showing how much grammar sits between the two.

BThe most careful measurement of the phenomenon comes from Nuala Beckett, whose group at the Ardmore Centre for Signed Language asked fluent signers and hearing non-signers to judge the same 1,100 signs between 2013 and 2017. Signers rated 43 per cent of the items as clearly motivated by the appearance or the use of the thing named, a figure far higher than most linguists had expected; hearing participants shown the same signs without context, however, arrived at the correct meaning for only 9 per cent. Beckett argues that resemblance in these languages is largely visible in hindsight, and that the property therefore cannot do the work the older accounts assigned to it. She also reports a constraint that she regards as more important than the figures: as a sign becomes grammatically productive, the pictorial detail is worn away, so that verbs which have acquired agreement are rated as less transparent than nouns of the same age. Resemblance, on her account, is tolerated only where the grammar has no use for the form.

CGideon Halloway, a developmental psychologist at Threave, accepts Beckett's ratings and disputes what has been built upon them. Her claim that resemblance is unhelpful, he maintains, rests on adults judging isolated items, which is not how any child meets a language. In his own study, twenty-six deaf children of deaf parents were followed for eighteen months, and the signs their parents recorded as first produced were, on his coding, more strongly motivated than the vocabulary as a whole by a margin of about eleven percentage points. Halloway is careful about what this shows. He concedes that the diaries were kept by parents who knew the purpose of the study, that a single language was involved, and that the effect had disappeared by the children's third year, which is consistent with Beckett's argument that resemblance matters least once grammar takes over. What the two disagree about, in the end, is not whether the property exists but whether it does any work.

DThe strongest evidence that these are grammatical systems rather than refined gesture comes from the way space is used. A signer establishes positions in the area in front of the body for the people under discussion, and thereafter verbs move between those positions, so that direction alone distinguishes who did what to whom. Hearing people asked to invent gestures for the same scenes place their hands consistently, but they do not reuse a position established several sentences earlier, which is precisely the step that turns a habit into a system. The rule is not universal in the way early accounts implied. Spatial agreement of this kind is elaborated most fully in languages with large communities of users, and appears in a reduced form, or not at all, in village languages confined to a few hundred people among whom deafness is hereditary. In one such village, first described in 1994, where one resident in twenty-five is deaf and hearing neighbours sign competently, the positions are used for emphasis but rarely for agreement, a difference that has been read, mistakenly, as evidence of simplicity. Beckett insists that the reduced form follows from the size of the community rather than from any poverty of the system.

EDescription has not kept pace with any of this. Of perhaps 150 sign languages now in use, fewer than forty have a published grammar of any length, and the recording of the rest is complicated by the absence of a written form, so that a corpus consists of video which must be annotated by hand at something like sixty times real duration. Motion capture, in use since 2009, has reduced the labour for certain purposes without solving the problem, since what has to be annotated is meaning rather than movement. Funding tends to follow description, which makes the gap self-perpetuating, and the varieties least documented are usually those with fewest advocates. Meanwhile the communities themselves are being reshaped by changes in education and in medicine, and the village varieties are the most exposed, precisely because the conditions that produced them are the least likely to recur. On the question that began the field, the fairest verdict is that both the old dismissal and the newer enthusiasm were built on one error, and that resemblance turns out to be a resource which each language exploits in its own way and then, as the grammar tightens, gradually declines to notice.