[Editor's note: Due to the font limitations of HTML, underlining on vowels will be used to represent a Polish hook (centered right hook), and a horizonatally barred l will be used to represent an l with a diagonal slash.]

This squib is intended as a contribution to the issue of what sorts of morphological building blocks and combinatory processes are possible in natural language.^* Fernald and Napoli (forthcoming) propose a model of lexical relatedness in American Sign Language (ASL) that is similar to the model proposed by Faltz (1999, 2000) for Navajo. However, these analyses reveal certain essential differences between ASL and Athabaskan.

The phonological analysis of ASL signs are generally assumed to involve four complex parameters (Hand Configuration (HC); Orientation (O); Place (P); Movement (M)), along with more subtle factors things secondary movement and facial expression. The study of ASL phonology has benefited from the work of many researchers. Brentari (1998) includes a recent survey of this body of work. Below are representations that Fernald and Napoli (forthcoming) assume for the ASL signs shown in the illustrations in the appendix.¹

(1)	FAMILY:	[HCF Ma Oa Pa]
	TEAM:	[HCT Ma Oa Pa]
	CLASS:	[HCC Ma Oa Pa]
	GROUP:	[HCG Ma Oa Pa]

From these representations it is evident that the signs in (1) differ only with respect to HC. In addition to having the other parameters in common, these signs have a great deal of meaning in common.

Fernald and Napoli (forthcoming) present a model of the ASL lexicon in which signs are built of "ion-morphs" by unification. An ion-morph is a triplet consisting of physical form, meaning, and a restriction on the form/meaning association. The ion-morph shown in (2) can unify with one of the ion-morphs in (3) to yield the signs represented in (1).

(2)	[x, Ma, Oa, Pa] = 'whole' where x {HCF, HCT, HCC, HCA, etc.}
	Note: two-handed simultaneous echo

(3)	[HCF, x] = 'family' where x {[Ma, Oa, Pa]}
	[HCT, x] = 'team' where x {[Ma, Oa, Pa]}
	[HCC, x] = 'class' where x {[Ma, Oa, Pa]}
	[HCA, x] = 'association' where x {[Ma, Oa, Pa]}
	etc.

The restriction is crucial since these forms have the meanings indicated only when they unify with compatible ion-morphs. Two ion-morphs are compatible just in case the form of each ion-morph is listed in the restriction of the other.

The examples shown thus far could possibly be analyzed as involving a root and affixes. The ion-morph in (2) carries specifications for the bulk of the parameters and it also carries much of the meaning of the fully formed signs. However, it is frequently the case with ASL signs that no ion-morphs can be identified as a root. Consider the representations below (and again see the appendix for illustrations of the signs):²

(4)	FATHER:	[HC5 Mb Ob Pb]		MOTHER:	[HC5 Mb Ob Pc]
	NEPHEW:	[HCN Mc Oc Pb]		NIECE:	[HCN Mc Oc Pc]
	GRANDFATHER:	[HC5 Mf Ob Pb]		GRANDMOTHER:	[HC5 Mf Ob Pc]
	FUTURE:	[HCB Mf Ob Pd]

These signs are composed from the ion-morphs shown below:

(5)	[x, Pb] = 'male' where x {[HC5, Mb, Ob], [HC3, Mb, Ob], [HCN, Mc, Oc]}
	[x, Pc] = 'female' where x {[HC5, Mb, Ob], [HC3, Mb, Ob], [HCN, Mc, Oc]}
	[HC5, Mb, Ob, x] = 'parent' where x {Pb, Pc}
	[HCN, Mc, Oc, x] = 'child of sibling' where x {Pb, Pc}
	[x, Mf, Ob] = 'on and on' where x {[HC5, Pb], [HC5, Pc], [HCB, Pd]}

In no way is it reasonable to any of these elements to be regarded as roots and for others to be considered affixes. Hence, Fernald and Napoli (forthcoming) used the term ion-morph to suggest that none of the morphological components of these signs are more central to forming the sign than any others. As Fernald and Napoli (forthcoming) show, ASL classifiers, in this analysis, are simply ion-morphs in which only the HC is specified:

(6)	ASL "Classifiers" (motion verbs)
	[HCw, Mx, Oy, Pz] = 'move'
	x is determined by properties of the referent of movement,
	y is determined by properties of the referent of the thing that moves with respect to the endpoints of movement,
	z is arbitrary referential points in the signing space,
	and w {3, 1 or inverted 2, bent 2, ...}, determined by properties of the referent of the thing that moves.

The building blocks of the ASL lexicon bear a striking resemblance to those of the Athabaskan languages. Faltz (1999, 2000) proposes that Navajo verbs are built by unifying a verb theme with one or more "satellites", the latter being a term Faltz borrows from Talmy (1975). A satellite, in Faltz's conception turns out to be remarkably like an ion-morph: it consists of a phonological form that is paired with a particular meaning just in case it combines with one of a certain set of verb bases. Below are examples of Navajo verbs with common roots³ but with different satellites:

(7)	[JOOL: non-compact matter] (Young 2000:4–5)
	'Aghaa' bikáá'adání biyaadéé' hááljool		'I took the wool out from under the table.'
	Sitsi' sitsiigháshchíín shá shik'iidííljool		'My daughter put my wig on for me.'
	Kin bikáádéé' 'aghaa' la' 'adááljool		'I tossed/dropped/lowered some wool from the roof.'
	Níyolgo shich'é'édáa'gi ch'il deeníní naajool		'A tumbleweed is tumbling around in my dooryard.'
	Tsiigháshchíín yishjolé—shílák'ee háájool		'The wig that I was carrying fell out of my hand—I dropped it unintentionally.'
	Sitsiigháshchíín tsásk'eh bikáa'gi dah shijool		'My wig is lying on the bed.'

So we see that a single root can combine with various satellites. Below are examples of a single satellite with phonological form ha that unifies with several different stems:⁴

(8)	ha- 'up, out' (position Ib) [biyaa 'under him/her/it']
	Biyaa haashheeh		'to take it (LPB) from under him/her/it.' (YM 1987:432)
	Biyaa haashjááh		'to take it (PlO1) from under him/her/it.' (YM 1987:433)
	Biyaa haashjiid		'to lug it (LUG) from under him/her/it.' (YM 1987:433)
	Biyaa haashjool		'to get it (NCM) from under him/her/it.' (YM 1987:433)
	Biyaa haashkaah		'to take it (OC) from under him/her/it.' (YM 1987:433)
	Haashkaal		'to chop it out' (YM 1987:433)
	Biyaa haash'á		'to send him/her out from under him/her/it.' (YM 1987:434)
	etc.

To see how Faltz's model works, consider the derivation of the verb in (9).

(9)	nídiilhaal 'you (sg) (are about to) club him/her.' (Faltz (1999:2))

The components of this verb minus inflection are shown in (10) and are derived by unification from the theme and satellite in (11).⁵ Faltz proposes creating structures like those in (10) (which he says correspond roughly to the traditional Athabaskan notion of verb base) from themes and satellites by unification: "to create a verb base, each field in [the theme] and [the satellite] is copied into the appropriate slot in the structures in [the verb base] and [the stem set]. As long as there is no conflict, the unification succeeds and a verb base results. In the case of a conflict, we expect to say that the elements cannot unify to create a verb base; but in certain cases, additional principles will have to be invoked" (1999:6).

(10)	stem-set:
		root:	GHAAL
		stem-aspect:	momentaneous
	classifier:		l
	conjugation-pattern:		YLV
	prefixes:		ná, d

(11)	theme:
		root:	GHAAL
		classifier:	l
	satellite:
		stem-aspect:	momentaneous
		conjugation-pattern:	YLV
		prefixes:	ná, d

In general, a lexical entry in the Navajo verb lexicon consists of an element that can be morphologically described as a verb base (as diagrammed in (12)) and derivationally described as the unification of a verb theme (as diagrammed in (14)) with one or more satellites (as diagrammed in (15)).

(12)	verb base:
		stem-set
		classifier
		conjugation-pattern
		prefixes
(13)	stem-set:
		root
		stem-aspect
(14)	theme:
		root
		classifier
(15)	satellite:
		stem-aspect
		conjugation-pattern
		prefixes

The Navajo lexicon has numerous satellites. Below are a few additional examples:

(16)	{Paa} + N + momentaneous	'give it to P'
	{ná, d} + LV + momentaneous	'pick it up'
	{ni} + N + momentaneous	'set it down'
	{} + S + neuter	'keep it'
	{'ahíI} + FLV + momentaneous	'mix it together'
	{yisdá} + simple + momentaneous	'rescue him/her'
	{na} + S + continuative	'carry around'
	{Pidá, d} + N + momentaneous	'cover P with it'

Unification alone does not make all the necessary restrictions on what satellites can compose with what themes. So, at least some satellites or some themes must also include a list of the elements with which they are compatible. This makes them formally very much like ASL ion-morphs because they pair form and meaning only when they combine with certain other morphological units.

There are, however, significant differences between the building blocks of ASL and the building blocks of Navajo verbs. First, while Navajo satellites may be phonologically null, a Navajo verb stem is complete phonetically in the sense that they are pronounceable, even though they are not fully formed words. ASL ion-morphs, on the other hand, are incomplete phonetically: if any of the parameters of a sign are unspecified, it is impossible to articulate the sign.

Second, as shown above, ASL ion-morphs can compose with each other to yield a full-fledged word, without composing with anything one would call a root or a stem. This is not the case in Athabaskan in which one morphological unit clearly is a stem. This is a major difference in the sorts of objects that are found in the two lexicons.

Faltz (1999) points out a significant difference between Navajo and Indo-European languages. To expose the difference, Faltz distinguishes "head level" morphemes (roots, stem-sets, and themes), which determine the choice of verb stem in a significant way, from "non-head level" morphemes (prefixes and satellites), which function like modifiers on the head levels. Faltz argues that "Many event types of the sort denoted by (say) simple Indo-European verbs cannot be expressed in Navajo by means of simple verbs. The reason is that such event types are not defined by virtue of the actions that carry them out, but rather by virtue of some other characteristic of the event" (Faltz 1999:8). Thus the meanings of the satellites in (16) would all be expressed in Indo-European languages by head level elements. Navajo differs from Indo-European languages in using non-head level elements to express these meanings.

Faltz proposes the following generalizations:

(17)	Cognitive levels associated with the major levels in the Navajo verb lexicon:
	verb base:	a fully specified event or state type
	theme:	a general class of action or state
	satellite:	variety of semantic types, including specific event-types definable without reference to the action performed to carry them out
	stem-set:	none
	root:	a very general notion of an action/state type, or group of these
	prefix:	varying

(18)	Cognitive types associated with grammatical/derivational categories of the Navajo verb lexicon:

	stem-aspect:	inherent aspect
	classifier:	transitivity
	conjugation-pattern:	???

And most significantly:

(19)	A Navajo theme NEVER denotes an event type.

In sum, Indo-European languages use head level elements to denote the event type. Athabaskan languages use non-head level elements (satellites). The head level elements express qualities of the verb's theme argument.

Now consider how ASL fits into this picture. ASL handling and motion verbs, like those in Athabaskan, are classificatory. Like Athabaskan, event types are realized as non-head level elements. But the case of ASL could not be otherwise: ASL verbs have no head-level elements whatsoever.

Appendix

	FAMILY	TEAM	CLASS	GROUP
	FATHER	MOTHER	GRANDFATHER	FUTURE

References

Brentari, Diane. 1998. A Prosodic Model of Sign Language Phonology. Cambridge, MA: MIT Press.

Faltz, Leonard. 1998. The Navajo Verb: a Grammar for Students and Scholars. Albuquerque: University of New Mexico Press.

Faltz, Leonard. 1999. The Structure of a Lexicon: Navajo (and other) Verbs. Paper presented at the West Coast Conference on Linguistics.

Faltz, Leonard. 2000. Some Notes on Derivational Relationships among Navajo Verbs. Papers in honor of Ken Hale. MIT Working Papers in Linguistics.

Fernald, Theodore B., and Donna Jo Napoli. forthcoming. Exploitation of Morphological Possibilities in Signed Languages. Sign Language and Linguistics.

McCarthy, John. 1981. A Prosodic Theory of Nonconcatenative Morphology. Linguistic Inquiry 12: 373–418.

Talmy, Leonard. 1975. Semantics and Syntax of Motion. Syntax and Semantics 4:181–238.

Young, Robert W. 2000. Navajo Verb Morphology: an overview. Albuquerque: University of New Mexico Press.

Endnotes

^* Thanks to Eva Furrow and Jane Ng for the ASL illustrations. [Back]

¹ The value for the HC parameter in each case corresponds to the value that hand configuration has in fingerspelling. The letter a used to represent the other parameters was chosen arbitrarily. [Back]

² Illustrations for FATHER, MOTHER, GRANDFATHER, and FUTURE are shown in the appendix. GRANDMOTHER is just like GRANDFATHER except for the initial hand position. NIECE and NEPHEW are like FATHER and MOTHER except that they use the N handshape. As with the examples in (1), the parameter values for HC correspond to fingerspelling values, but the letters representing the other values were selected arbitrarily. [Back]

³ Navajo verbal morphology is magnificently complex, and I must gloss over details here. A verb base consists of the verb root and specifications for stem aspect and various prefixes including the so-called classifier. Different stems of the same root are represented in (7): some examples have the l classifier and others have Ø. See Faltz (1997, 1999, 2000), Young (2000) for details. [Back]

⁴ The verb root in each of these examples is classificatory. The abbreviations LPB, PlO1, LUG, NCM, and OC are used by Young (2000) to indicate certain characteristics of the object being transported. [Back]

⁵ Faltz writes, "The internal lexical structure of a verb as viewed by the morphology does not have the same architecture as the internal lexical structure of a verb as viewed by the semantics: morphologically, the root determines the stem-set (together with the stem-aspect category), whereas derivationally the root determines the theme (together with the classifier). This mismatch is not peculiar to the Navajo language—we expect to find similar mismatches in the lexica of all languages" (1999:5). Later he adds, "If we imagine that the Navajo verb lexicon consists in part of a list of themes and a list of satellites, it would appear that a lexical verb is constructed by choosing a theme and a satellite and combining them as shown in (11). But to determine the actual forms of the verb, a representation such as the one in (10) is better, since that representation corresponds directly to the actual morphology of the verb forms" (1999:6). [Back]