The target article “Reported speech as a dedicated syntactic domain: Typological arguments and observations” by Spronck and Nikitina constitutes an ambitious attempt at defining the core properties of reported speech constructions from a typological point of view. Even though the amount of descriptive work in sign languages is still limited when compared to spoken languages, there exists a significant amount of research on this topic. It would have been particularly interesting to incorporate those results into the empirical basis on which the proposal is built. This could also have influenced the choice of label for the phenomenon discussed in their Section 1.2: a more encompassing choice of label such as “reported discourse” might have included automatically both languages in the oral-auditory modality and in the visual-gestural modality, and the reference to other types of reported discourse, such as thought, might have been more transparent. However, beyond the terminological choices, what is particularly interesting from the proposed typological approach is that the evidence provided by sign languages in the domain of linguistic reports provides further support for some of the claims made in the article.
From early on, sign language research identified that in linguistic reports a strategy was used whereby the signer adopts the “role” of the reported illocutionary agent (for an overview, see Lillo-Martin 2012). This strategy is known as “role shift”, and descriptively it is flagged by the following general characteristics for the sign languages described so far:
- Body lean towards the referential location associated with the reported signer, sometimes only marked with head lean.
- Facial expression associated with the reported signer.
- Eyegaze break with the actual interlocutor at the beginning of the reported stretch.
- Interpretation:Person, time and space parameters are shifted to the reported context and interpreted accordingly.
Role shift is the most genuine way sign languages have to encode reports. The overt markings can combine, but eyegaze change seems to be required (Herrmann & Steinbach 2010). The phenomenon has received different characterizations that compare it to reported discourse in spoken languages. One of the most recurrent questions in the relevant works is whether role shift can be equated to quotation in the strict sense, that is, mentioning the utterance that someone else produced in a different context, thus corresponding to example (1) in the target article. Relying on spoken language categories, it seems well established that sign languages can also produce typical indirect discourse, as example (2) in the target article. Both examples are repeated here for simplicity of exposition.
John said: “Look, there is marmalade here!”
John says that there was a typhoon yesterday.
Interestingly enough, role shift as described above only marks structures such as (1) corresponding to direct discourse, and never indirect discourse of the type in (2). This is illustrated in examples (3) and (4), respectively, in Catalan Sign Language (LSC). In (3) the tier above the report stands for all the nonmanual markers of role shift that identify overtly the reported utterance anchored to the main clause subject, namely the source of the report. Example (4) features simply a case of an indirect report realized as an embedded complement clause of a verb of saying which is not marked as role shift. In Figure 1 it can be observed how role shift markers are coarticulated with the embedded subject pronoun IX-1 in (3), in contrast to the third person pronoun IX-3 in (4) in Figure 2, without role shift (the raised eyebrows observed are a marker of topic). 1
annai 3-say-2 ix-1ifed-up lose+++ (LSC)
‘Anna told you that she was fed up with losing so often/ Anna told you: ‘I’m fed up with losing so often.’
annai 3-say-1 ix-3ifed-up lose+++ (LSC)
‘Anna told me that she was fed up with losing so often.’
The target article questions such a strict divide of reported discourse into direct and indirect speech reports (Section 2.6). Sign language evidence seems to support this position in general, as we will see below. However, an important general asymmetry between the two types of cases in sign languages is that indirect discourse needs to be introduced by some linguistic means (an expression of the matrix of the report, M), otherwise the report (R) is interpreted as an utterance of the actual signer. By contrast, role shift overtly marks R in sign languages and R often stands alone, especially in narrative contexts, as far as it is clear to which referent the report must be anchored, in a similar way to the Nyulnyul case (8) discussed in the target article.
Next to the fact that indirect speech is not marked by role shift in sign languages, research has shown that role shift structures cannot be strictly classified as pure instances of quotation as a whole (Quer 2016). This has been determined on the basis of different properties, some of which we will mention here briefly. First, in LSC lexical verbs of saying introducing role shift are in principle ambiguous between a strictly quotational and a non-quotational reading (see for instance the LSC example (3) above), but certain lexical introducers such as author, declare, voice or say1-sentence unambiguously introduce literal quotes.
annaiexplain say-1 sentence ix-1ibrother man 3-ignore-1(LSC)
“Anna told me: ‘My brother ignores me.’”
Second, in LSC and in German Sign Language (DGS), place indexicals like here appearing in R can be interpreted not in the displaced context of the report but in the main context of utterance. Thus, in example (6), uttered in Barcelona, the indexical here is interpreted as “here, in Barcelona” by default.
ixamadridmmoment joanithink ix-1istudy finish hereb (LSC)
‘When he was in Madrid, Joan thought he would finish his study in Barcelona.’
However, it has been shown that in other sign languages like American (ASL) or French Sign Language (LSF) this is not a possibility (Schlenker 2017a), which supports the idea in the target article that the grammar of particular languages is regulating the interpretation of indexicals in R, with significant cross-linguistic variation.
Third, although wh-question formation is typically forbidden out of a quotation in a language like English, it turns out to be possible in ASL (Schlenker 2017a).
in la who ix-a [johna] say ix-1 will meet here who(ASL)
‘In LA, who did John/he say he would meet there [in LA]?’
Again, this kind of wh-extraction out of a role shift fragment is not possible in another sign language like LSC, further supporting the observation that the grammatical properties of particular (sign) languages determine the syntactic and semantic properties of R in reported discourse. 2
Sign languages can provide interesting evidence to examine the claim that “reported speech constructions are conditioned by grammar” (Section 2.8). In isolation, this statement seems rather trivial, but it points in fact to an important empirical observation: gesture and prosodic cues typically accompany R in reported speech. The authors of the target article subscribe to the idea that these are multimodal properties that are often coarticulated with reports, but do not constitute proper marking (however, they remain agnostic as to whether this can be the case in some languages). This point connects directly with the proper characterization of role shift in sign languages.
In the visual-gestural modality, gesture and expressive prosody are even harder to disentangle from signing proper because they are all realized in the same medium, and it is often far from trivial to distinguish the linguistic part of the signal from the paralinguistic, expressive or communicative part. Role shift itself bears on the surface the reproduction of the gestural and the facial expression of the reported agent. The relevant literature has mostly focused on its use to report utterances and has used different labels to this end (constructed dialogue, as in Metzger 1995; attitude ascription, as in Quer 2013, or attitude role shift as in Schlenker 2017b, among others). Nevertheless, it is clear that in signed discourse role shift, which necessarily marks the linguistic report, also flags the imitation of actions or gestures carried out by the individual the report is about. This type of role shift occurs systematically intertwined with the utterances or thoughts of the reported agent, in a richer and more structured fashion than in co-speech gesture. For researchers that distinguish between the two uses, the latter one is labelled constructed action, as in Metzger (1995), or action role shift as in Schlenker (2017b). 3 Other researchers, though, do not make such a clear-cut distinction and treat role shift structures as a single phenomenon that reports on utterances and actions with the same means, and call it constructed action (Cormier et al. 2013) or personal transfer (Cuxac 2000). From this perspective, we observe that the debate addressed in Section 2.8 of the target article reproduces itself along very similar lines in the study of role shift in sign languages. At bottom lies the question whether gesture and expressive prosody have to be accounted for by the linguistic system or not, and the answers vary as to how the linguistic system is defined. Roughly speaking, the more restrictive view focuses on the study of grammar and lexicon as the core language faculty and leaves outside “paralinguistic” or co-expressive signals like co-speech, co-sign gesture or expressive prosody, albeit recognizing that they align with the linguistic signal and closely interface with it. The more encompassing view that studies language as the broader communicative ability of humans broadens the focus to include those signals, which may be not determined by grammar in the strict sense but are part of the linguistic system. This is an extremely interesting and thriving area of research, and more extensive research on role shift in sign languages should contribute directly to the debate.
The authors of the target article put the emphasis on reported speech being approached as a “dedicated syntactic domain” or a “crosslinguistic syntactic category”. They correctly emphasize the morphosyntactic variation attested in the realization of R and the optionality of the occurrence of M. However, the definition they offer in (31) (Section 3.1) points towards a more semantic-pragmatic nature of the defining components. Without having to renounce the study of the grammatical variation in the ways R is realized and how it is constructed with M, it looks safer to conclude that reported discourse is a discourse phenomenon identifiable on semantic-pragmatic grounds in discourse generally, but encoded by grammatical means. This does not mean that certain incarnations of reported discourse cannot be analyzed in terms of syntactic subordination of R to M, with all the consequences thereof, as in (2). This is, for instance, a requirement for certain role shift structures where a quantifier binds a pronominal element in R, as in the LSC example (8):
nooneisay ix-1iagr-1 scared darkness (LSC)
‘Noone says he’s scared of darkness.’
The fact that role shift (and reported discourse more generally) is a discourse phenomenon does not exclude the possibility of treating it with grammatical means. This approach is proposed for instance in Quer (2005, 2011), where role shift is analyzed as the realization of a context-shifting operator, or in Herrmann and Steinbach (2010), where it is seen as an agreement phenomenon. In both cases the role shift nonmanual markers are characterized as the overt grammatical markers of the structures under examination. Grammar can thus be argued to also be at the core of such a cross-sentential phenomenon as role shift/reported discourse: as we have seen above, the particular grammars of individual languages determine diverging properties in a phenomenon like role shift that looks very similar on the surface. Only further in-depth work into these matters in more sign languages will allow for a better understanding of the interactions between the strictly linguistic properties determined by the grammar and the gestural and prosodic expressive signals intertwined with the report. Advancement in this domain will obviously help disentangle the same issues arising for the analysis of reported discourse in spoken languages.
From these few observations we can safely reach the (by now trivial) conclusion that in order to achieve a truly cross-linguistic characterization of a natural language phenomenon, it is not only advisable but also necessary to take the evidence from sign languages into account. Some of the claims made in the target article are clearly supported by sign language evidence, thus strengthening the cross-modality of the proposed criteria and generalizations; others might be reconsidered on the basis of such additional evidence. A lot of descriptive and analytical work lies ahead for sign language linguists, but seeing those results incorporated into current description, analysis and theorizing of human languages will boost the required work in sign language research.
This contribution has been made possible thanks to the SIGN-HUB project, which has received funding from the European Union’s Horizon 2020 research and innovation program under grant agreement No 693349, as well as to the Spanish Ministry of Economy and FEDER Funds (FFI2015-68594-P), and to the Government of the Generalitat de Catalunya-AGAUR (2017 SGR 1478).
Cormier, Kearsy, Sandra Smith & Martine Zwets. 2013. Framing constructed action in British sign language narratives. Journal of Pragmatics 55. 119–139.
Cuxac, Christian. 2000. La langue des signes française. Les voies de l’iconicité. Paris: Ophrys.
Herrmann, Annika & Markus Steinbach. 2010. Eine neue Perspektive auf Role Shift in deutscher Gebärdensprache (DGS). Perspektivwechsel als nichtmanuelles Kongruenzphänomen. Das Zeichen 84. 112–118.
Lillo-Martin, D. 2012. Utterance reports and constructed action in sign and spoken languages. In Roland Pfau, Markus Steinbach & Bencie Woll (eds.), Sign languages. An international handbook, 365–387. Berlin: Mouton.
Metzger, Melanie. 1995. Constructed dialogue and constructed action in American sign language. In Ceil Lucas (ed.), Sociolinguistics in deaf communities, 255–271. Washington, D.C.: Gallaudet University Press.
Quer, Josep. 2005. Context shift and indexical variables in sign languages. In E. Georgala & J. Howell (eds.), Proceedings from semantics and linguistic theory, vol. 15, 152–168. Ithaca, NY: CLC Publications.
Quer, Josep. 2011. Reporting and quoting in signed discourse. In Elke Brendel, Jörg Meibauer & Markus Steinbach (eds.), Understanding quotation, 277–302. Berlin, New York: Mouton de Gruyter (Mouton Series in Pragmatics).
Quer, Josep. 2013. Attitude ascriptions in sign languages and role shift. In Leah C. Geer (ed.), Proceedings of the 13th meeting of the Texas linguistics society, 12–28. Austin: Texas Linguistics Forum.
Quer, Josep. 2016. Reporting with and without role shift: Sign language strategies of complementation. In Annika Herrmann, Roland Pfau & Markus Steinbach (eds.), A matter of complexity: Subordination in sign languages, 204–230. Berlin: Mouton de Gruyter.
Schlenker, Philippe. 2017a. Super monsters I: Attitude and action role shift in sign language. Semantics and Pragmatics 10(9). doi:
Schlenker, Philippe. 2017b. Super monsters II: Role shift, iconicity and quotation in sign language. Semantics and Pragmatics 10(12). 1–67. .
I follow the usual glossing conventions in the sign language literature, according to which manual signs are represented by the capitalized word corresponding to the gloss of the sign (the most general translation into a spoken language). The scope of nonmanual markings is represented with a line that spreads over the manual material with which it is coarticulated. The relevant abbreviations for the purposes of this paper are the following: #-VERB-# (verb agreeing with subject and object; the number before the verb refers to the grammatical person of the former and the one after the verb refers to the latter); IXa (locative index pointing to locus a); IX-# (pronominal index; the number corresponds to person); rs (role shift); t (topic marking); +++ (reduplication of the sign). The referential indices i, j, a, etc. link the first person role in role shift fragments to the intended author of the reported utterance.
Other dependencies like long-distance NPI licensing in R by a negative element in M are excluded in ASL. For details, see Schlenker (2017b).