Emergence or Grammaticalization? The Case of Negation in Kata Kolok

Round 1

Reviewer 1 Report

Preliminary comments

First, I must report that after reading no more than a page of this manuscript I realized I had read it several years ago. It is a slightly edited down version (c. 12,957 words vs c. 14,659 words) of Lutzenberger, H. (2017). Negation in Kata Kolok: Grammaticalization throughout three generations of signers. Universiteit van Amsterdam, Master's Thesis, Amsterdam, c. 14,659 words (supervised by Roland Pfau, Vadim Kimmelman, Connie de Vos). It is at least 4 years since the original was completed. The final accepted manuscript should state this and there should also be a brief explanatory statement describing key differences between the two versions. I remember that I thought the 2017 version was very good and noteworthy in its methodology and overall findings on previous research on Kata Kolok (KK). It provided corpus-based data on negation strategies in KK that presented a significantly different picture of negation in the language than had been reported in the literature up to 2017, as well as raises issues about the manual vs non-manual typological types re negation in signed languages. This difference in conclusions about headshaking seemed related to the nature of the data — naturalistic conversations in KK, rather than primarily elicited material as in Marsaja (2008).

Second, I needed to do several library searches as part of preparing my review for this paper. I needed to check some text citations that were not listed in the references. The missing references are: Coppola and Senghas 2010: Dahl 1979; de Vos 2015; Dryer 2005; Franklin et al 2011; Frishberg 1975; Gökgöz 2011; Johnston 2018 ; Lucas et al. 2001; Motamedi et al. 2019; Nyst 2012; Pfau & Quer 2002 (or is this a mistake for Pfau & Quer 2007?); Schembri et al. 2009; Supalla 2001; Zeshan and de Vos 2012. I also found there were items listed in the reference list that are NOT cited in the text: Benitez-Quiroz et al 2016; Biberauer & Zeijlstra 2012; Branchini 2021; Lutzenberger in preparation; Mietsamo 2007; Mietsamo 2013; Pfau & Quer 2007 (mistake? see comment above); Spruijt et al in preparation; Tang 2006. This is the most poorly referenced paper I have ever reviewed, and I have reviewed many over the years. At most one finds one or two such mistakes, not 22! This did not create a good impression.

General comments

I can fully understand that the author or the authors (henceforth ‘authors’) wish that the basic findings of the thesis be published so that it reaches a wider audience: it deserves to be. It is a pity that this had not been done soon after the thesis was passed and awarded. In publishing a thesis as a journal article, especially after 4 years have passed, one expects there to be further new work or at least reinterpretation or contextualization the original findings in the light of any subsequent scholarly work in the field. The authors have only slightly updated the 2017 original in this 2021 version. They have added about 14 new references, but only two (Johnston 2018, & Motamedi et al. 2019) post-date the 2017 original. Only Johnston 2018 is of direct relevance to the this manuscript. The Johnston citation refers to: Johnston, T. (2018). A corpus-based study of the role of headshaking in negation in Auslan (Australian Sign Language): Implications for signed language typology. Linguistic Typology, 22, 185-231. It reports in further detail data first reported in (which it cites) Filipczak, J., Johnston, T., Kuder, A., Mostowski, P., & Rutkowski, P. (2015). Between non-manual gesture and grammar: Is headshake a negation marker in Polish Sign Language (PJM) and Australian Sign Language (Auslan)? Paper presented at the Workshop: Nonmanuals at the Gesture Sign Interface (NaGSI), Goettingen, Germany.

(Motamedi et al 2019 describe experiments that describe how gestures used by hearing non-signers change over time when asked to use them communicatively without speech including transmitting/teaching them to other people.)

Johnston 2018 observed that in the corpus-based studies of negation and headshaking in Auslan, NGT (he cites in the work of Oomen & Pfau 2017) and KK (he cites the work of Lutzenberger 2017) raised questions about the typology of negation proposed by Zeshan (2004, 2006) as it applied to each of those three sign languages. He also questioned the empirical basis of the typology and whether it accurately reflects negation in signed languages. Johnston 2018 is thus relevant to the overview of negation in signed languages (in Section 3.2 pp 5-6) of this manuscript because presents data and observations that post date the 2017 thesis. The data presented for KK in Lutzenberger 2017 and in this manuscript are similar to Johnston 2018 for Auslan but with interesting differences. A comprehensive overview of negation in signed languages should include more details on the data and conclusion from this corpus-based study, not just from Oomen & Pfau (2017). More positively there is a second citation of Johnston 2018 at line 647 and at line 651 the authors observe: “Our results also [like Oomen & Pfau 2017 & Johnston 2018, cited previously at line 647] suggest that prior classifications based on elicited data may have to be re-evaluated.”

Specific textual comments:

I did not see any typos or simple errors. 

lines 126-140

The desire to restrict the analysis to standard negation is understandable (essentially to enable comparison of this basic operation in KK with that described in the linguistic typology literature). The authors are aware that this may not be so straightforward in signed languages partly because of the difficulty sometimes in deciding clause/sentence boundaries and the grammatical function class of constituents therein (lines 135-140). This may mean that ‘standard negation’ may actually not be easy to define for some signed languages or that they may have more than one ‘basic’ clause negator (cf lines 141-144).

lines 182-184

However, this is extremely rare. In most languages with tonal negation, it is reported that there is usually a negative particle or morpheme present in the clause in addition to the negative tone.

line 242-249

Given the findings of Oomen & Pfau re NGT, Lutzenberger 2017 (i.e., and this very paper), and Johnston 2018, it may be important to re-consider how previous research is described and it may be prudent to be less definitive, i.e., does previous research really “reveal different spreading options for headshake..”? or rather is it more accurate to say “it has been claimed there there are different spreading options for headshake...”. Similarly, consider tempering “These combinatorial restrictions strongly suggest that the headshake is not just a co-speech gesture” with something like “If confirmed by the analysis of corpus data, these combinatorial restrictions would strongly suggest.....”

This would simply parallel the treatment of pre-corpus studies elsewhere, e.g., line 313 where the authors write “Marsaja (2008, p194) claims that neg is mandatory in all negative utterances....”, and not “reveals”, “finds”, “shows” etc., or lines 423-424 “Marsaja (2008) claimed to be the most common one.”

line 250 & footnote 4

As explained above, to give an accurate picture of the literature on negation in signed languages to date the large corpus-based study of headshaking and negation in a signed language cited in this footnote really should be described in more detail. In other words, it should include discussion of other non-manuals insofar as they have already been mentioned in the literature in the context of negation. As it stands, it is a very selective account. Johnston 2018 draws attention to non-manuals during negation in addition to or instead of headshaking (nodding and negative facial expressions, incl. mouth gestures) and points out the impact this may have on the interpretation of headshaking and other non-manuals as formal markers of negation (as opposed to other types of symbolic behaviour that can co-occur with denial, negative semantics, and grammatical negation). Similarly Zeshan’s original typology (2004, 2006) also discusses negative facial expressions (i.e., excluding head shaking or backwards head tilting). She does not considered them to be formal markers of negation.

lines 306-308

The reason for not reporting on not.yet and finish does not make sense. In lines 325-327, it is said the study aims “to establish which KK marker is the main one”, but it seems that in lines 306-308 this has already been assumed.

lines 326-327

If (ii) is a real point of enquiry, then it should be explained why the function of headshaking across KK discourse generally is not thought to be relevant, and not considered. Also, by using the term ‘marker’ (rather than, say, ‘facial expression’ or ‘non-manual expression’) is one not assuming what one is trying to establish?

lines 372-385

The qualifications and caveats given in this paragraph are well made and important to consider. Indeed, they are so important that given the size of the dataset, as reported, that generalizations regarding language change and grammaticalization, for example, are difficult to make. Ironically, the authors may even agree with this given their concluding paragraph to this paper.

lines 410-411 (i.e., Figure 2)

It would be extremely helpful if the category “non-manual only” was specified for headshake, tongue protrusion (tp) , or both. From my reading of the figures and the surrounding text I calculate that in Gen III perhaps all 13 are headshakes, in Gen IV all 3 are tongue protrusion, and in Gen V all 3 are also tp, i.e., at most 6 out of 162. (It turns out from line 511 that there are only 3, as the other 3 are both headshake and tp.) Details of discourse context is needed to know if negative semantics associated with this expression could not explain its presence and the (assumed) clause/sentence negation attributed to these three tokens. Also with respect to figure 2, it would be best to label “manual-only” as simply NEG to align the description with all the negative clauses/sentences marked with NEG (irrespective of any other non-manual activity).

lines 436-437

What is really important from the point of view of headshaking is not so much topic as the discourse context: what is the signer doing (answering question, denying something, expressing contrary information, merely making an incidental observation, etc.)? Here and elsewhere the manuscript does not take account of the issue of the discourse situation and headshaking (both in negated and non-negated clauses), apart from a brief comment that “subject matter had no impact” on negation strategies. Johnston 2018 showed that in the Auslan corpus, at least, discourse situation (answering a question, denying an asserted state of affairs, rather than simply making a negate statement in the absence of any strong presuppositions to the contrary) has an effect on the likelihood of headshaking, nodding, or negative facial expressions (of which tongue protrusion is also another possible one in Auslan) occurring during negation.

lines 479-481 (example 8c)

Re previous comment, without context there is no way for the reader to know if something like ‘No, I’m not angry’ is also a possible paraphrase of the signed utterance. In that case, it is possible headshaking could be seen to be anaphoric and interactive: ‘No (you are wrong to say/imply ‘You are angry’), I am not angry.’ and not strictly speaking marking negation of the clause uttered. There may be nothing to stop such a non-manual spreading over the entire utterance, especially if it consists of only a few signs.

Section 6.2.2.

Tongue protrusion (often with wrinkled nose) is a very common facial expression in many (all?) human groups that indexes/instantiates displeasure, disgust, rejection etc. Tp can, in turn, symbolize a general negative stance (i.e., displeasure, disgust, rejection, disappointment etc.). Perhaps its use with KK signers is congruent with this general idea, rather than it does not occur in the surrounding community at all? What is the context of the examples? Also (as mentioned above) there seems to be only a very small number of examples in the dataset where tp is claimed to be the only possible negator morpheme present. On line 511 it is stated there are 3. Though they may have prompted a negative translation into Balinese/Indonesian/English is it correct to attribute such a grammatical structure to the KK utterances? Are other explanations also possible? More contextualization and more examples seems to be needed. The subsequent paragraph at lines 527-557 does suggest negative evaluation really is quite central to tp.

Lines 575-576

Since it is based on a corpus-based sample of negated clauses in Auslan and reports the same thing, it would be relevant to also mention Johnston 2018 at this point.

Lines 624-626

Do the authors mean by “all sign languages” all the sign languages listed in Table 3, or all sign languages generally speaking? If the second interpretation is intended, Johnston 2018 clearly shows that the negative particle is predominately pre-predicate in Auslan (62%) so a correction would be needed. If the first interpretation is intended then this should be made clear.

Section 7.2

The speculations re language change in this section are interesting, but I feel too much is being made of very limited data, be they of negation patterns generally or just the non-manuals. (I’m not sure if both or one is meant.) This seems to be conceded in lines 666-667, but it is immediately followed by a generalization (lines 667-668) which is tenuous. The caveats that then immediately follow in that paragraph (listed one to four) suggest very good reasons why even if the generalization could be sustained by the data, that the interpretation of it, as offered, is open to question.

Also, it seems too much space is dedicated to Pfau’s 2015 suggestion re Jespersen’s cycle and grammaticalization of headshake and the primacy of the manual negative particles. The authors suggest (reasonably) that manual negative particles and non-manual gestures were in the emergence of the KK system from the very beginning, even though “this proposal is not grounded in empirical evidence”. Critical self-awareness is welcome, but we do know that headshaking and negative facial expressions are universal co-speech gestures, often accompanying negation. It is unlikely that any stage of signed language and spoken language evolution lacks manual or non-manual gestures,

lines 750 onwards, including Conclusion

The authors make a very accurate summary of the limits of the study and list factors and additional data types that are crucial to evaluating claims made in this study. It is unfortunate that in the four years since the original 2017 version, no attempt seems to have been made to do some of this for this paper.

Author Response

Please see attachment.

Note that the footnote numbers provided in the revision table correspond to the version with tracked changes (non-sequential due to changes made). 

Author Response File: Author Response.pdf

Reviewer 2 Report

The paper investigates negation in a sign language used in Bali, Kata Kolok. They present interesting results with regard to the distinction between manual and non-manual markers and also with regard to specific features such as tongue protrusion and headshake. The results are based on the analysis of 1.73 hours of spontaneous language data of six native signers.

 

Overall, this paper is very compelling and written in a coherent and logical way. For a study which addresses language typology issues and considers aspects of language change, the analysis of six native signers as ‘a corpus-based approach’ is a bit generous. The results of two signers from each generation should be taken with caution and it is highly problematic to conduct statistical analyses with less than 15 individuals. This is discussed by the author in the paper but it would be useful to see the degree of individual variation presented in the results too and to see if this differs across generations.

 

The paper should be contextualized in the wider emergence literature more generally, including work about grammaticalization of relative clauses by Dachkovsky and more general findings about young sign languages and the use of the body. One important point which seems to be overlooked is the fact that in the early stages of sign languages, there has been a trend, noted in several studies, for some sign language features (e.g., classifiers) to start on the body and move to the hands. This might actually account for one of the results found in this study – the fact that the manual negator is used more by recent generations. I suggest that this literature is added and then discussed in the Discussion.

 

Minor comments:

In the Abstract: I believe Mesch – Mesh, if it’s Kate Mesh.

Section 2.1 Emergence – should be expanded. Include Dachkovsky’s work and Aronoff et al. (2003), Engberg-Pedersen (1993), and Kegl et al., (1999). Finally, add in information about the contribution of the body in language emergence: Sandler (2012), Meir and Sandler (2019), Sandler (2017), Dachkovsky et al. (2018).

 

Section 2.2: the first two paragraphs seem redundant. I suggest dedicating this section to grammaticalization?

 

Page 12: describe the degree of individual variation in the dataset and across generations.

 

References:

• Aronoff, , Meir, I., Padden, C. & Sandler, W. (2003). Classifier complexes and morphology in two sign languages. K. Emmorey (Ed.), Perspectives on classifiers in signed languages, Lawrence Erlbaum Assoc, Mahwah, NJ, pp. 53-84
• Dachkovsky, S. (2020). From a demonstrative to a relative clause marker: Grammaticalization of pointing signs in Israeli Sign Language. Sign Language & Linguistics, 23(1/2), 142-170.
• Dachkovsky, , Stamp, R. & Sandler, W. (2018). Constructing Complexity in a Young Sign Language. Front. Psychol. (2018), 10.3389/fpsyg.2018.02202
• Engberg-Pedersen, E. (1993). Space in Danish Sign Language. The semantics and morphosyntax of the use of space in a visual language. International studies on sign language research and communication of the deaf. J. Linguist., 19. (406).
• Kegl, , Senghas, A., Coppola, M. (1999). Creation through contact: sign language emergence and sign language change in Nicaragua Language Creation and Language Change: Creolization, Diachrony, and Development, MIT Press, Cambridge, MA, pp. 179-237.
• Meir, & Sandler, W. (2019). Variation and conventionalization in young sign languages. E. Doron, M. Malka Rappaport Hovav, Y. Reshef, M. Taube (Eds.), Linguistic Contact, Continuity and Change in the Genesis of Modern Hebrew, John Benjamins, Amsterdam, pp. 337-363.
• Sandler, W. (2012). Dedicated gestures in the emergence of sign language. Gesture, 12 (3), pp. 265-307
• Sandler, W. (2017). The challenge of sign language phonology. Annu. Rev. Linguist. (2017), pp. 43-63

Author Response

Please see attachment.

Note that the footnote numbers provided in the revision table correspond to the version with tracked changes (non-sequential due to changes made). 

Author Response File: Author Response.pdf

Reviewer 3 Report

I have read this paper carefully twice, once when I first received it (although unfortunately due to unforeseen circumstances, I could neither annotate it nor write my review at the time), and a second time over the past few days when I have been able to comment directly on the ms. (a commented version of which I will find a way to attach to this review).  On the first reading, I liked the paper a lot and thought it well worth publishing, although there were inchoate worries in the back of my mind.  

On the more detailed second reading, I still think both the data and the argument sound and a worthwhile contribution to science.  My doubts have become a bit clearer, however, and I will build them into my suggestions.  I have thus been able with confidence to fill out the checkboxes on your review form, and I will make more substantive comments here. 

Against the backdrop of a typologically oriented comparative study of sign languages, the paper breaks new ground by presenting a small corpus-based study of negation in a small, village sign language, Kata Kolok (KK), which has developed in a Balinese village with a high incidence of hereditary deafness (and an associated deaf-oriented culture).  The specific data set has been assembled from a much larger corpus of KK by selecting  selected “naturalistic” interactions dyads of signers from three “generations” of the language, which were then (partially?) (re-?)transcribed with special emphasis on the formal marking of “negation.”  I put the word in scare quotes because, as is well known, negation is a complex notion in natural languages, despite many attempts to reduce it to an apparently simple “logical” notion that is defined in terms of propositional truth value. (One obvious omission in the paper’s bibliography is, indeed, Larry Horn’s foundational and voluminous work on the subject.) There are indirect references to this complexity in the author’s (I presume there are multiple authors, but for simplicity I will use the singular here)  use of such formulations as “standard negation” (which is, according to the abstract, part of “the typological landscape” of sign languages, because of the prominent role of ‘clausal negation’—another rubric—within such a “landscape”), as well as later rubrics like “logical (i.e., sentential) negation,” which is linked explicitly to the “standard” kind, which is in turn linked explicitly to “polarity” and “truth values”.  A slightly different classification has to do with (syntactic) scope: does negation apply to “sentences” or “constituents” (which are further elaborated formally—linked to specific parts of speech)—that is one kind of scope. Another has to do with some sort of pragmatic or speech act negative “function” (as the author again recognizes laterally, contrasting “affirmative” to “negative” sentences or polarities, and contrasting the notion of “affirmation” with, for example, a negative imperative).  Later in section 5.2, things are still further muddled by a whole series of notionally different kinds of “other negative forms” (including “negative interjections, negative existentials, etc.” that get coded but then presumably have to be filtered from the results, in some way.  Finally, it is not until the penultimate sentence of the entire paper that the author comes clean with a much fuller catalogue of “other negative forms,” only then revealing the complexity of what has been taken as an almost unproblematic primitive up until then.  “Negative” is just one of the areas where typology seems to be preeminent in framing the argument, rather than, say, facts of KK.  Moreover, as the author points out, the criteria for typological classification of sign languages with respect to negation are actually quite different from those applied to spoken languages. And, as it turns out by the end of the article, the author thinks they may well be wrong, something that we don’t learn until fairly late in the piece (the first paragraph of section 7).

There is nothing inherently wrong with this sort of mixed categorization, except that it seems to me the wrong way to work when considering a young and little studied language.  Why start with typological categories (based often on wholly different languages) rather than simply forms in the target language?

The authors argue that KK has three signs, both “manual” and “non-manual,” that seem to have to do with negation (or which, in formal terms, they would like to consider as a group on some grounds).  The proper procedure, as it seems to me, would be then to present the distributional facts of these forms in the selected corpus, their co-occurrences, their formal and syntactic variation and apparent classes of functions, and work from there to a set of empirical results about their distribution, across utterance types and speakers, and then—if desired—hazard some typological proposals.  In fact, this is exactly what I think the paper DOES, substantively. But the order of presentation, and the flip-flopping between typological assertions based on other languages, and the KK facts, obscures for me the basic arguments in the paper:

1. That KK (apparently) adopted into the sign language a series of gestural emblems that have “negative meanings” of different kinds in the surrounding speaking/hearing communities in the course of its development.
2. That these include both manual and “non manual” gestures (most notably a hand wave and a head shake, and—a new contribution here—a protruding tongue which seems, both syntactically and semantically, to link to the other two).
3. That these are conjointly used by KK signers, contrary to previous claims, and that each can also occur by itself.
4. That one of these “non-manual” signs—because of its articulation on the head—can “spread” in the sense that it can occur simultaneously with other signed elements, and perhaps its “spreadability” is increasing (and, as a result, so is the efficiency of the language in signaling its syntactic scope) among younger signers.

You will find other specific comments directly on the marked up PDF I have submitted with this review.  My specific suggestions for the author are:

Switch up the order of presentation to make clear where you intend to END UP with this paper, rather than (pretend?) to build up a typological argument which in the end you intend to question.  Then the garden-paths introduced in the (for me) premature application of typological claims can be quickly sealed off. 

Elaborate more on how a co-speech gesture may become incorporated into a sign language, so that it becomes subject to the pressure of possible grammaticalization. (And say more about how that process of grammaticalization is, classically, meant to work and how it might have worked in this case.)  That is, at least consider the potential mechanisms involved, both in emerging grammar and over generations of signers.

Finally, say more about your transcripts, which I find uninformative in some respects—for example, the precise temporal unfolding of different synchronous and simultaneous elements in the signing.  Is it impossible to show any illustrations of the examples from the corpus? In any case, the abbreviations and orthographic conventions used in transcripts need to be spelled out.   

Comments for author File: Comments.pdf

Author Response

Please see attachment.

Note that the footnote numbers provided in the revision table correspond to the version with tracked changes (non-sequential due to changes made). 

Author Response File: Author Response.pdf