Review Reports - The Functions of Prosody in Action Formation in Australian Greek Talk-in-Interaction

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The paper presents research on the very important issue of prosody in talk-in-interaction and highlights the importance of the analysis of prosodic features in bilingual conversations. More specifically, it offers an enlightening analysis of the way prosodic features serve as contextualization cues that frame the different functions of discourse markers and repetitions delivered in a code that constitutes a departure from the base-language/code of the interaction. The questions posed are timely and interesting.

All in all, the paper exhibits sound documentaation of the proposed thesis and is well written. It could be published as it is.

Author Response

Thank you very much for taking the time to review this manuscript. Thank you for your positive comments.

Reviewer 2 Report

Comments and Suggestions for Authors

Review of

“The functions of prosody in action formation in Australian Greek talk-in-interaction”

Summary

The current paper investigates the prosodic structure of English-Greek bilingual discourse markers and bilingual repetition with naturalistic data gathered from the first-generation members of the Greek diasporic community in Cairns, Far North Queensland (Australia). The authors (hereafter As, in plural) convincingly show that in both case studies the phonetic structure assumes a well-defined pragmatic function. Concerning the discourse markers, As show that bilinguals use prosodic information for displaying their stance and carrying out specific actions in conversation. As for the bilingual repetitions, As reveal that bilingual speakers employ changes in duration and intensity to signal language change and, they therefore render the switch boundary as participant-related.

Overall evaluation

As overall finding that prosodic cues are employed meaningfully in insertions and bilingual repetitions in bilingual speech, and the fact that they landed on this conclusion following a well-chosen and well-explained methodology, and with data from an understudied diasporic community deserve admiration. I am quite sympathetic to the manuscript, and I strongly recommend its publication. What few minor comments and corrections I provide below do not alter this decision.

Minor points

[1] ll 390-392: “When speakers insert English discourse markers in Australian Greek, they seem to draw on prosodic resources, which are locally meaningful and subjected to intra-speaker variation.”

While the first conclusion has been made clear throughout the text, the latter, namely, that these resources “are subject to intra-speaker variation”, as far as I see, is not properly discussed in the paper.

[2] ll. 30-37: “The Greek community in Cairns consists of three generations: […] This study focuses on first-generation Greeks (50-90 years old) […]”

I am not sure if the three-way classification takes into consideration only those who migrated after WWII or those who settled in Australia at the end of the 19^th c. as well.

[3] ll. 59-60: “Discourse markers (also known as ‘pragmatic markers’ or ‘pragmatic particles’) are small, uninflected words or phrases that are devoid of syntax […]”

Please notice that a growing body of work, mostly couched within the framework of Chomksyan-generative framework, discusses that discourse, hence discourse markers, can be syntacticized. See Haegeman and Hill(2013) and references therein for earlier work.

[4] l. 73: Furthermore,

[5] ll. 151-155: Is there a principled reason why these three participants were chosen among the 11 first-generation speakers?

[6] section 2.2.: are the recordings stored in wav format?

[7] ll. 339-341: ‘[…] the code switch found in the second iteration is not interactionally motivated, as it does not serve to highlight the referents’ twin relationship, but rather displays the speaker’s preference for Greek as the appropriate language of interaction.”

Does this then mean that the speaker has evaluated using the word ‘twins’ is not well taken by the address and thus felt the urge to ‘correct’ herself? (Sorry, I am just trying to understand what must have possibly motivated the switch)

Just one more question, is the agreement marker on the Greek word (dhidima) OK in this context from a monolingual’s perspective?

Reference

Haegeman, Liliane and Virginia Hill. 2013. Syntacticization of discourse. In Syntax and its limits, ed. By Raffaella Folli. Oxford: OUP.

Comments on the Quality of English Language

The quality of English of the manuscript seems quite fine to me; however, given that I am not a native speaker of English, I cannot state my opinion firmly. I would still kindly ask As to reread their manuscript carefully for typos and grammar mistakes.

Author Response

Thank you very much for taking the time to review this manuscript. Please find the detailed responses below and the corresponding revisions/corrections highlighted in red colour in the re-submitted file.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

This is a clearly written paper on the linkages between prosody and action formation—something that has not very much been investigated in the past. Furthermore, the authors’ specific way of investigating this linkage in the context of code switching is even more novel—let alone the choice of the specific languages to be investigated. However, I have some concerns with the paper in its current form which point to a need of revision before the paper can be published.

1. The literature review on prosody and action formation in lines 103-123 involves references to very many studies on the topic. What I am missing, however, is a more careful selection of, and focus on, those studies that are specifically relevant from the perspective of those very actions that your analysis sheds new light on. What is known about these actions, so far, and how can your study contribute to the understanding of the formations of these specific actions? Given the limitations of space, I would therefore suggest that you might drop references to some of these studies and focus more closely on those that are particularly relevant to your argument.

2. In a similar vein, I am missing a more thorough discussion from the final section of the paper (Discussion and concluding remarks). How do your findings relate to prior literature on prosody and action formation in CA and IL. Now the paper’s contribution to this body of research remains open.

3. It is not very clear how the reference to the three different types of code switching in lines 47-52 is relevant to your study. Specifically, the point about there possibly being code switches that “do not carry any locally defined meanings” (line 52) is surprising, given the CA and IL approach taken in the paper (cf. Harvey Sacks’s notion of there being “order at all points”).

4. The representation of the pitch contours would be much more informative if you would indicate the pitch range of the given speaker both in the table in which you introduce the three speakers and in the PRAAT pictures that you show. At the moment, a reference to a speaker’s “default pitch register (line 226) is left without substantiation, while a reference “to a higher F0 (around 377 Hz)” is not really meaningful without information of the speaker’s pitch range.

As has been demonstrated in the study by Lennes et al. 2015, in many languages, pitch distributions are skewed to the right. This means that the speakers generally stay around their typical pitch level, while sometimes deviate from that level for specific interactional purposes. Some researchers use the minimum and maximum pitch as indications of pitch range, although I don’t recommend it (see below). The median is a more reliable indication of the speaker’s pitch range (which can also be obtained from PRAAT). But what is an even better indicator of pitch range is the speaker-specific mode, which is not affected by non-modal voice and the selected analysis parameters. (Unfortunately, to my knowledge, the mode cannot be directly extracted from PRAAT, but the pitch information need to be considered in Excel etc.) People’s pitch range has been shown to be between approximately 6 semitones below and 12 semitones above their mode, which allows one to estimate the pitch speaker's pitch range quite nicely. (As shown by Lennes and colleagues, the pitch mode converges quickly to a rather stable level and the overall standard deviation drops under 1 semitone even after the analysis of only 34 s of the speech by the given participant. This means that even a very short amount of speech by the given speaker can allow the researcher to estimate the speaker's pitch range, without this estimation relying on whether or not extreme pitch levels are included into the sample.)

Then, if you know the speaker-specific mode (i.e., the typical pitch level that the given speaker uses most of the time), you may indicate it with a dotted line in PRAAT pictures (see e.g., Korpela, et al. 2023).

A further possibility to increase the meaningfulness of PRAAT pitch contours is to define the lower and upper end of the scale used in the picture with reference to each speaker’s mode (i.e., the most typical pitch level, defined in Hz), so that the pitch range (as pointed out above) between 6 semitones below and 12 semitones above their mode is also the size of the window in each picture. The currently used window of 50-500 Hz is very broad, in which case the pitch contours appear relatively flat, making the subtle pitch movements difficult to observe.

Furthermore, I think that the size of the window should be the same for each PRAAT picture (as defines in semitones, not in Hz!), but that the upper and lower limits of the pictures could vary, depending on what the pitch range of the given speaker is.

5. Lines 312-313. Please, conclude the section by briefly summarizing the findings regarding the use of prosody in influencing the actions implemented by code-switched discourse markers.

6. Lines 364-366: Please, also conclude this section by summarizing the ways in which the changes in prosodic configurations shape the function of the repeat. (As the title of the paper suggest that the study is about “action formation” try to be as precise with this as possible!)

References:

Korpela, R., et al. (2023). Prosodic linking in apology sequences in Finnish elementary school mediations. Journal of Pragmatics, 214, 54-72.

Lennes, M., et al. (2015). Comparing pitch distributions using Praat and R. Phonetician.

Author Response

Author Response File: Author Response.pdf