Professional Documents
Culture Documents
1. Overview
This chapter describes how native users of American Sign Language
(ASL), a visual-manual language of the deaf in the United States, use hands
and body as articulators to mark noun classifiers in ASL verbs of motion
and location. The chapter describes in some detail both the physical
articulators and the shapes they adopt fornoun classification (i.e., the phonol
ogy of ASL classifiers), and the morphological organization of these clas
sifiers.
The main point of this review is to show that the classifier system in
ASL is strikingly similar to classifier systems in spoken languages; and, more
over, that this is true despite the potential in ASL for a rather different type
of organization. As will be discussed below, a language in the visual-manual
mode obviously has a very different set of resources at its disposal, from
which to build a morphology, than does an auditory-vocal language. The
objective of this chapter is to show how a visual-manual language organizes
these rather different resources for classifying nouns, and yet still shares the
same types and linguistic functions of noun classification found in spoken
languages.
objects in the real world. In contrast to this claim, Siipalla (1982, in press)
and Newport (1981, 1982) have demonstrated on the basis of both linguistic
analysis and psycholinguistic data that ASL verbs of motion, like words of
spoken language, are composed of combinations of discrete morphemes;
and that the morphological parameters and grammatically possible values
along these parameters are like those found in spoken languages of the world.
ASL and spoken languages do differ, however, in two ways: in ASL, but
not in spoken languages, each of the morphemes is sometimes transparently
(or translucently) related to its meaning; and the morphemes tend to combine
simultaneously, rather than sequentially. These latter two characteristics give
some ASL verbs the appearance, though not the actual structure, of an
analogue representation of real-world objects and motions.
These and related linguistic analyses (Frishberg, 1975; Kegl & Wilbur,
1976 ; Supalla, 1978) have revealed that ASL verbs of motion include classifier
morphemes: that is, morphemes that vary in form as a function of certain
categorical properties of the associated noun. For example, the handshapes
of ASL verbs may vary as a function of whether the associated noun is a
two֊ vs. four-legged object; or they may vary as a function of whether the
associated noun is long and thin vs. broad and flat. See particularly, Supalla
(1982, in press) for evidence that these morphemes are indeed classifiers
(and not, for example, depictions of objects).
In the present chapter, I will present the range of classifier categories
found thus far in ASL. These categories will be taken primarily from my
own work (most especially, Supalla, 1978, 1982, & in press), although there
is some additional work on ASL classifier morphemes which I will cite. My
claim here is that, despite the apparent iconicity of classifier morphemes in ASL,
and despite the fact that these morphemes could therefore potentially depict
an infinite number of distinct real-world objects, ASL has developed only
those types of classification that are found in spoken languages of the world.
Throughout the chapter I will note similarities between the two kinds of
languages in types of classification and in the particular categories found
within each type. These similarities suggest rather strongly that the recurrent
categories of classification found in spoken languages as well as signed lan
guages are not due to any particular restrictions of the auditory-vocal mode,
or to the limited resources available within this mode for morphological
distinctions.
Before beginning this presentation, I will first sketch the overall
structure of the ASL verb of motion.
THE CLASSIFIER SYSTEM IN AMERICAN SIGN LANGUAGE 183
as well. For example, much of the morphology of verbs of motion and location
in Navajo (including some of the classifier system) is said to be frozen and
unanalyzed. In ASL, however, the morphology of verbs of motion is currently
highly productive in native signing, with frozen structures confined only to
individual, commonly used lexical items. In the present chapter, I will focus
only on the productive, morphemic noun agreement markers.
tion of the noun. 4) The fourth way of classifying nouns is when the hand
or some other part of the signer's body is used to represent a bodypart in
the referent (that is, a bodypart classifier). The body articulator is involved
here as a location for the bodypart. The mimetic representation, in this case,
is appropriate only when a human hand is referred to. Otherwise, if some
other bodypart (e.g., head, mouth or foot) is referred to, a visual-geometric
representation would be used instead. 5) A final type of classification is by
the type of instrument (either a hand or a mechanical instrument) which acts
on the object. An instrument classifier may either be derived from mimetic
representation (in the case where the instrument is a hand manipulating the
object) or from visual-geometric representation of the instrument (in the
case where the instrument is mechanical). In either case, the object being
manipulated is not directly referred to. In sum, there are at least five separate
groups of classifiers, each group derived independently through different
morphophonological processes.
To illustrate the independence of the derivational processes for these
classifier groups, the morphophonological organization will be compared
across the five different classifier groups enumerated above: size-and-shape
specifiers, semantic classifiers, body classifiers, and instrumental classifiers.
First, I will start with the derivations of the size-and-shape specifiers. Signific
ant phonological and semantic features will be introduced at each of several
derivational levels. Some examples of phonological and semantic contrasts
will be presented. Similar descriptions will then be given of the derivational
processes for the other five groups of classifiers. In the subsequent section
(Section 2.6) I will then discuss several additional morphological processes,
in which morphemes may be affixed to a classifier to represent qualities of
the noun such as its consistency and physical integrity, or for more than one
noun, their quantity, and relative locations or status. The final section (Sec
tion 2.7) then shows how the classifier system can be manipulated so that
new classifiers can be synthesized to refer to new classes of nouns.
which agree with the noun in various aspects of its size and shape. Each
SASS actually consists not of a single handshape morpheme, but of a group
of simultaneous hand-part morphemes: each finger as well as the thumb and
forearm is possible morpheme which can combine in specifiable ways to
form a handshape (Supalla, 1978; in press). These hand-part morphemes
each classify a different aspect or dimension of the visual-geometric structure
of the noun referent. For example, one hand-part marks whether the referent
is a dot (i.e., 0-dimensional) or a one-or-more dimensional object; another
hand-part marks whether it is straight or round in shape; and so on. These
hand-parts together form a static SASS. In addition, certain handshapes can
have a limited set of movement morphemes added to represent further dimen
sions of the object, by a kind of "tracing" or outlining of the dimension in
space. The latter I have called tracing SASSes (Supalla, 1978; in press).
The process of combining hand parts into handshapes is a productive
system in which parts must be combined in a fixed order. If only one mor
pheme is marked on the unmarked hand, the specified part of the hand
becomes an independent morpheme. But there are some hand parts (e.g.,
+ middle finger) that are not independent morphemes, as they can only be
added to the hand already marked with another hand-part morpheme. Such
morphemes are considered bound morphemes which can be added only in
later derivational levels.
The morphophonological structure of static SASSes is shown in Figure
3. The SASSes included here are from my dialect. Each level of derivations
of the static SASSes is described successively in the following subsections.
Other types of SASSes like the tracing SASSes are then discussed afterwards.
sifiers has nothing to do with legs of the object. This group includes three
different types of classifiers, each referring to a certain function as well as
the orientation of the object (i.e., freely maneuverable, horizontal object
for the first subgroup, freely maneuverable, vertical object for the second
subgroup, and fixed columnar object for the third subgroup). Classifiers
within each subgroup vary in number and arrangement of finger/thumb/
forearm extensions that reflect the shape of the referent noun.
Furthermore, all of these groups of classifiers are specialized in move
ment. Legged object classifiers can be marked with movement morphemes
representing the manner of locomotion of the object along a path. Maneuv
erable object classifiers would require a different set of movement mor
phemes representing the object's course along a path. On the other hand,
the fixed object classifiers refer to unmaneuverable objects fixed to their
bases and, thus, are restricted to local movements like waving or shaking
back and forth.
Despite the differences in handshape and movement, all of these clas
sifiers in the two groups are phonologically more simple than the SASSES
and seem not to be derived from morphemic handparts as are the SASSes.
They are constructed of simple handshapes with fingers and thumb extended
flat from the hand. The center of the body of the referent object is represented
by the unmarked hand and each marked finger or thumb represents an exten
sion protuding from the body of the object. The separation of extendedness
concepts from shape and size concepts is not unique to ASL as Denny (1976a)
shows that the classifiers in spoken languages make this distinction.
vertical human beings and other upright animals. The hand may be further
derived where each other finger and thumb extension would function as a
morpheme in the derivational process to quantify the noun. Since the seman
tic function of such classifiers is different, this process is further discussed in
a later section covering the system of number classifiers in ASL.
The other underlying form has the hand oriented vertically on its side.
There are several classifiers which are phonologically contrasted by exten
sions of the hand parts. One classifier has all four fingers extended forward.
This classifier refers to saliently two-dimensional vertical objects (e.g., a
fish). In contrast, when the vertical object has salient individual projections,
fingers are added to the hand with each finger representing a projection from
the body of the object. The vehicle classifier, described earlier in this paper,
originated as a classifier for a sailing vessel with projecting masts.
There are several restrictions in using the body classifier system. First,
the body classifier can be used only when the noun referent is animate.
Second, when a body classifier is incorporated into a verb, it can mark only
one referent object (although various locations on the body can be marked
to refer to various attributes of the same referent object). Third, the reference
scale must be consistent when body classifiers and classifier handshapes are
combined, in other words, the relative sizes of the referent nouns constrain
the combination of these classifiers. Fourth, if the whole body is used to
refer to an agent carrying out an action, the body classifier cannot be com
bined with a movement path across space. Verbs involving body articulators
can only express manners of movement (but not paths of movement), whereas
verbs involving hand articulators can express either or both of these aspects
of motion (see Talmy, 1972, for related restrictions in verbs of motion in
spoken languages).
hand tightly together while the hand moves ; this combination refers to smooth
textures. The second member of this pair has the hand instead moving in an
irregular fashion to refer to rough textures.
3. Summary.
This chapter has overviewed the various types of classifiers thus far
discovered within verbs of motion and location in American Sign Language.
Overall, the types of classification found in ASL, as well as the particular
categories found within the classes, are like those found in spoken languages
(See Adams, Delancey, Denny and Creider, Erbaugh, Payne in this volume,
Erbaugh 1984). Moreover, this is true despite the potential in a visual-manual
language for a more analogue, depictive representation of objects which
would in principle permit rather different or more detailed classification than
might be found in a spoken language. These similarities thus underline the
role of basic human linguistic and categorization principles, and not the
particular characteristics of the input-output modality, in the organization
of classifier systems.
Acknowledgements
This reasearch was supported in part by NIH Grant No. NS 16878 to
Elissa Լ. Newport and Ted Supalla, Department of Psychology, University
of Illinois at Champaign-Urbana. I want to express my gratitude to Jenny
Singleton and Elissa Newport for their valuable help with the use of the
English language in the writing of this paper, and to Elissa Newport for
helpful discussion of ASL classifiers. All figures were drawn by myself; Fig
ures 3,10 and 13 were adapted, with permission, from American Sign Lan
guage andshape Cards, produced by Dawn Sign Press.
REFERENCES
Table 1
English Translations of ASL Verbs of Motion in Figure 1
Figure 5. HOUSE
208 TED SUPPALLA
Figure 7. HIT-IN-THE-EYE
THE CLASSIFIER SYSTEM IN AMERICAN SIGN LANGUAGE 209