Stokoe (1960). Sign Language Structure Silver Spring Maryland: Linstok Press.
Why are there phones? Why doesn't one word of English have clicks and another have ejectives and a third have high back unrounded vowels?
It would be a lot to remember: ease of encoding.
Probably even more important: ease of decoding. Acoustic signals are complex.
There are a lot of meanings out there. And a lot of noise in the world of signals.
With no phones, the problem of distinguishing among 1000 morphemes would be the problem of remembering and distinguishing 1000 discrete distinctions among continuously varying signals like that spectrogram.
With phones, we reduce the problem to making 30 or 60 distinctions reliably within that continuous stream of signals, still hard, as the difficulty of building speech recognition systems has shown us, but way easier.
Moral: The problem of meaning recognition needs to be broken up into parts: distinguish among a small set of signals and associate sequences of those signals with meanings.
This is called duality of patterning. There are two systems of patterns: phones and morphemes. The phone system has fewer contrasts. By weaving together sequences of phones we are able to make many more distinctions in the system of morphemes.
Something similar goes on when we move from morphemes to sentences. By weaving together sequences of morphemes we are able to make many more distinctions among meanings than we make with the morphemes alone.
Another way of putting it: Duality of patterning is what allows linguistic systems to be as complex as they are.
Stokoe's work changed the way people looked at ASL.
Previously people thought a sign was something completely different from a word in spoken language.
After Stokoe's work (further pursued in collaboration with Casterline and Croneberg), it became clear that sign languages had duality of patterning as well:
In fact, all sign languages did. In fact all sign systems had meaningless repeatable subcomponents. Call these formational elements:
Why? For the same reason spoken languages do, probably. Having a continuous stream of 1000 signs with no formational elements would be as difficult to decode as a language without phones.
However there are some interesting lessons as well.
Notice that the formational elements of a sign do not have to be sequenced the way phones are. The handshape and motion can be simultaneous.
One of the interesting features of signing is that the medium allows independent elements to be simultaneous. And the system of formational elements exploits this.
The English spelling signs (one sign per letter of the English alphabet) are an entirely sequenced system. Native signers can spell out quite quickly and often do so for a variety of reasons (when speaking to beginning signers, for example), but they find extended spelling much less natural and efficient than signing.
However there are numerous sequenced features of signs as well. Action verbs that entail a spatial end point are characterized by a "hold" manner of movement at the end of the sign. Those with no spatial endpoint are executed with continuous movement at the end (Suppalla and Newport 1978)
What would be an example a formational element that can make a meaningful difference in language, but can also be simultaneous with other formational elements?
Thirty years ago when William Stokoe first published the first piece on ASL structure, the task he set for himself was to take sign-sized units and analyze them for their component phonological aspects (Stokoe 1960, Stokoe et al. 1965)> At that ime he concluded that there were three phonological components: the shape of the hand(s) when articulating a sign, which he called DEZ's, the location(s) on the body where the sign is articulated, which he called TABS; and the movement(s) of the hands in the course of changing locations or handshapes, which he referred to as SIGS. Stokoe purposely avoided terms that already had a use in linguistics, choosing CHEREME (from Greek xeir 'hand') as the superordinate term for these phonological components of signs, even though they were equivalent to phonemes in the American structuralist sense (as in e.g. Hockett 1958).
Some fairly random citations:
Constraint on hand shape: A morpheme may contain either the finger group or the finger group index-plus-middle-finger but never both.
The sequence: all-five-fingers + index-plus-middle-finger is forbidden.
Spoken language: Assimilation is one sound changing one of its features to be more like another sound nearby.
[ih n ] + [ k r eh d ax b ax l] = [ih ng k r eh d ax b ax l] ALVEOLAR + VELAR = VELAR VELARn assimilates to ng because the ng is pronounced in the same place (velar) as the following k sound.
ASL example (Fog. 10.2 p. 347). Assimilation ASL Browser [Communication Technology Lab: Michigan State University]
SLEEP + SUNRISE = OVERSLEEP 5 handshape C handshape 5 handshape C-handshape closed up closed up toward face + parallel to body parallel to face + parallel to facetoward face and parallel to face are like phonological features like place of articulation. Note that motion of head also assimilates. SLEEP involves motion downward of head, but SUNRISE does not. Head motion eleiminated of greatly curtailed in OVERSLEEP.
What is prosody in speech? Perceived features:
Sound stream features realizing prosody (what a phonetician would measure)
Sign language features for realizing prosody:
Example from Israeli Sign Language (ISL), F10.3, p. 348
Productive processes for forming new words. Knowing the meaning of the first sentences, we can deduce a lot about the meaning of the second:
There is also inflectional morphology, the -s ending on walks, the -ed ending on walked. And Swahili agreement:
Will sign language have anything like this?
The phenomenon: hand shapes that stand for classes of things combine with different types and manners of movement. Reminiscent of classifiers in Navaho and Asian languages.
Specific hand shapes classify objects, then combine with a special class of other words to describe the motion or the location of those objects:
The term iconic is used for signs that have some salient feature that resembles what they signify.
Onomatopoeia is a special case of iconicity. Onomatopoietic words are words that describe sounds or things that have a salient and sound something like what they are describing.
By and large Onomatopoeia is the main way in which spoken language signs can be iconic.
But sign languages have the added possibility of visual iconocity and this introduces some interesting possibilities:
On the other hand, sometimes the iconicity is more elusive, and often not present at all.
An interesting mix of iconicity and other features is in language about motion: