**6. Inflection**

*6.1. Inflection in Signed Languages*

An often-noticed contrast between signed languages and creoles is that while creoles have little or no inflectional morphology, signed languages are rich in it (cf. Aronoff et al. 2005).

First, most established signed languages have a class of verbs referred to in much of the literature as agreeing verbs (Padden 1988; Meir 2002; Lillo-Martin and Meier 2011), mostly verbs of transfer (literally and metaphorically) such as *give*, *send*, *take*, *help*, and *tell*, which have affixes indexed to the verb's arguments. Verb agreemen<sup>t</sup> in ASL is exemplified in Figure 3. The phenomenon differs in detail from sign language to sign language, but it is present in many established sign languages that have been studied. Meir (2012) documents the emergence of this kind of inflection in Israeli Sign Language, in which such verbs come to be rendered from and to points in space referring to the arguments. Emergent sign languages may not have agreeing verbs at the outset (Padden et al. 2010; Meir and Sandler 2008, p. 87); however, Rathmann and Mathur (2008) demonstrate that once established, this kind of marking becomes more complex over time.

**Figure 3.** Some examples of verb agreemen<sup>t</sup> in ASL. (**a**) I-GIVE-YOU, (**b**) I-GIVE-HIM, (**c**) (I-)GIVE-ALL. With permission from Carol Padden. (Spatial agreemen<sup>t</sup> with subjects other than first person, not pictured, also occurs regularly in ASL and other sign languages).

Second, signed languages develop a type of classifiers that manifest "depiction" verbs, of motion and location (cf. Emmorey 2013). That they belong to a discrete set which can occur redundantly with the specific nouns they refer to indicates their status as agreemen<sup>t</sup> inflection (Supalla 1996). The motions and locations that attach to these classifiers elaborate the meaning (cf. Meir and Sandler 2008, p. 111 for a useful cataloguing of classifiers in

Israeli Sign Language). Signed languages also indicate aspect inflectionally as shown in Figure 2 above.

#### *6.2. Inflection in Creoles*

The pidginization process eliminates all or most inflectional affixation, and it only reemerges in creoles very slowly. In some, sustained contact with a source language preserves a small amount of inflection, or allows it to be borrowed over time, often reanalyzing its behavior and function. For example, in Mauritian Creole, the French distinction between finite and infinitive verb stems is preserved as short and long forms derived from them, respectively, but reanalyzed; the long form carries what can be analyzed as an affix, with the occurrence of the forms determined by syntax, as shown in Table 2.

**Table 2.** Short and long verb forms in Mauritian Creole French.


Otherwise, inflection in creoles tends to occur in highly proscribed contexts. In Santome Portuguese Creole, *ba* "go" occurs as *be* when followed by an adjunct (*E ba ke* "He went home;" *E be d'ai* "He went from here" (APiCS)). In Saramaccan, with the same verb, "go," the imperfective proclitic *tá* occurs as *nán*-: *Mi tá wáka* "I am walking;" *Mi nángó* "I am going."

Saramaccan also has what can be analyzed as an object agreemen<sup>t</sup> marker, in the specific context of shared object serial verb constructions. In the sentence below, the low tones of "turtle" are fixed and thus do not participate in the rightward spread of high tone through the sentence. However, the high tone spread "jumps" over this object and alights upon the first syllable of the second verb *kulé*, which in citation form has a high tone only on its second syllable. Thus the tonal spread is a kind of object agreement.


#### *6.3. Implications for the Language Faculty*

The relative richness of inflection in signed languages makes it clear that it is an overgeneralization to stipulate that emergen<sup>t</sup> languages must be low on inflection. However, modality is the reason for the contrast between spoken and signed languages here.

For one, in signed languages inflection can be readily indicated iconically, via hand position or movement, or in the case of shape classifiers, by lexical signs recruited as inflections (Meir et al. 2010). In spoken language, inflection most readily emerges from a long-term process: the grammaticalization of lexical items. Also, as Poizner and Tallal (1987) noted, while visual processing of language is ill-suited to linear processing as rapid as what is possible in spoken language, "Signed languages have the potential for multiple channels for encoding grammatical information: face, head, torso, eyes, and various joints of the two arms can realize morphemically distinct information simultaneously" (cf. also Aronoff et al. 2005; Sandler 2018).

Language processed through the eye, then, develops inflection readily upon emergence; language processed through the ear does not. As such, inflection in emergen<sup>t</sup> spoken languages (i.e., creoles) is unexpected; in signed languages, what would be unexpected is its absence. However, Aronoff et al. (2005) note the relevance of the difference between simultaneous and sequential morphology. The latter, developed via grammaticalization of erstwhile unbound morphemes, is much less common in signed languages, with the reason for this being their youth. In this sense, signed languages parallel creoles, as they so often

do. (However, just as creoles are usually not completely devoid of new affixation, Polish Sign Language has developed several grammaticalized affixes marking negation, degrees of time, and "not yet" (Tomaszewski and E ´zlakowki 2021a, 2021b).)

## **7. Derivational Compositionality**

McWhorter (1998, 2012) argues that creoles are the world's only languages which combine three features:


We might ask the extent to which signed languages, as new languages, conform to this prototype.

As seen above, they do not, in terms of inflectional morphology. Also, obviously, the tonal aspect is irrelevant to signed languages. However, in terms of derivation, signed languages and creoles appear to pattern similarly.

#### *7.1. Derivational Compositionality in Creoles*

Derivational processes leave a cline of compositionality in older languages, such as in English, where we can identify four degrees:


In creole languages, born recently of pidgins, there are always cases of unpredictability in derivation, as no language could exist without them given the realities of culture and the vagaries of labelling. However, not enough time has passed for the emergence of cases of Level 3, where elements of a derived word have lost their synchronic meaning.

While generally overlooked in grammatical descriptions, cases like this are typical of languages that have existed for countless millennia (i.e., most human languages). They occur not only in combinations of roots with derivational prefixes, but in compounding, as in these Mandarin cases (Packard 2000, p. 222), given in Table 3:

**Table 3.** Analyzable but opaque compounds in Mandarin.


Creoles that have co-existed with their lexifiers borrow Level 3 cases from them, as in French creoles (explaining data regarding the prefixes *de*- and *re*- adduced by DeGraff 2005). However, in creoles that have not co-existed with their lexifiers, Level 3 cases are rare to nonexistent. For example, in Saramaccan there are only cases of Level 2— institutionalizations—rather than Level 3 (McWhorter 2013), as shown in Table 4. This is due to Saramaccan's emergence from a pidgin just some centuries ago.

**Table 4.** Level Two compounding in Saramaccan Creole.


#### *7.2. Derivational Compositionality in Signed Languages*

Compositional derivational morphology occurs in signed languages, but is limited (Aronoff et al. 2005; Sandler and Lillo-Martin 2006). However, compounding is very common in signed languages; e.g., compounds were 40% of the lexicon of the signed language of Providence Island in the Caribbean. However, studies sugges<sup>t</sup> that signed language compounds are of the Level 2 type rather than Level 3.

Emergent compounds have predictable meanings, such as ASL's BLUEˆSPOT "bruise," FACEˆNEW "stranger," LOOKˆSTRONG "resemble," REDˆFLOW "blood," and SEEˆMAYBE "check" (Klima and Bellugi 1979; Valli et al. 2011, p. 68; Sutton-Spence and Woll 1999, p. 102). Then, "frozen" compound signs develop, in which the meaning is less predictable, and the phonology differs from that which would express the elements in simple combination (Liddell and Johnson 1986), for example, BELIEVE from THINKˆMARRY in ASL, shown in Figure 4. However, signers often remain aware of the meaning of the elements within these frozen signs (Brennan 1990; Pfau et al. 2012b, pp. 171–72). In emerging signed languages, compounding often occurs productively, if erratically, on the fly.

**Figure 4.** ASL compound. The constituents (**a**) THINK and (**b**) MARRY, and the compound (**c**) BELIEVE. Images courtesy of the Sign Language Research Lab, University of Haifa.

This is equivalent to Level 2 compounding in spoken languages. Thus signed languages appear to conform to the Creole Prototype in this regard.

## **8. Conclusions**

The goal of this exploration has not been to merely show that creoles and signed languages have much in common. For one, the case that signed languages are emergen<sup>t</sup> languages would seem unexceptionable. Then, while the case that creoles are products of emergence rather than simply mixture is less universally accepted, the traditional pidginto-creole life cycle is empirically documented for some creoles and reconstructed for others with argumentation thus far unaddressed by critics in its details (cf. McWhorter 2018).

Thus, we would expect signed languages and creoles to share many features. The goal in this paper is to examine whether, on the basis of these similarities (as well as the dissimilarities), these two kinds of language can shed light on the nature of emergen<sup>t</sup> human language in general.

The conclusions to draw from the findings here sugges<sup>t</sup> that:


In sum, when language is emerging, it prioritizes marking the novelty of information (confirming Scott-Phillips 2015); is readily recursive (contra Everett 2005); favors the manner of action (aspect) over the time of action (tense); develops inflection readily only in a visual, as opposed to aural, mode; and develops derivational opacity only as the result of drift over long periods of time.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The author declares no conflict of interest.
