Previous Article in Journal
Colloquialization Processes in the 20th Century: The Role of Discourse Markers in the Evolution of Sports Announcer Talk in Peninsular Spanish
Previous Article in Special Issue
Is Negation Negative? (And a Discussion of Negative Concord in SOV Languages)
 
 
Article
Peer-Review Record

The Discourse Function of Differential Object Marking in Turkish

Languages 2025, 10(7), 173; https://doi.org/10.3390/languages10070173
by Klaus von Heusinger * and Haydar Batuhan Yıldız
Reviewer 1: Anonymous
Reviewer 2: Anonymous
Reviewer 3: Anonymous
Reviewer 4:
Languages 2025, 10(7), 173; https://doi.org/10.3390/languages10070173
Submission received: 13 May 2024 / Revised: 24 June 2025 / Accepted: 1 July 2025 / Published: 18 July 2025
(This article belongs to the Special Issue Theoretical Studies on Turkic Languages)

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This study presents three empirical studies that test the forward-looking function of DOM in Turkish and shows in two out of three studies that objects with and without DOM are not interpreted differently in experimental settings in terms of their forward-looking functions. The observation contrasts with the corpus study where human objects with and without DOM have different forward-looking patterns. The study presents a fourth study on the objects with the indefinite article bir, that display a clear forward-looking function.

DOM has been one of the most widely studied topics in Turkish formal linguistics. As the authors state, as well, it is mostly discussed in terms of its backward-looking or d-linking properties in the literature. In that sense, the present study is one of those rare studies that looks at it from a new perspective. It is also one of those rare studies that provide empirical data to support theoretical assumptions. Because of these properties, it would be a valuable contribution to the field in general and the audience of the Languages journal in particular. However, I cannot recommend its publication at this time, in its current form, because of the problems I explain in detail below.

Theoretical assumptions

1-      A major problem regarding the literature review and the discussion of theoretical assumptions is that, DOM is discussed within the context of d-linking, specificity or presupposionality and to some extent affectedness with famous examples from Enç and other literature. However, along with these examples, there are other examples in the literature where accusative marked indefinite objects are not d-linked. There are also empirical results that show that objects with DOM and without DOM are not different in terms of their interpretations under certain conditions, with certain types of verbs etc. (the manuscript cites these studies but does not discuss them within this context). Indefinite objects with DOM can be utterred “out-of-the-blue” in Zidani-EroÄŸlu’s terms. So it is very important for the purpose of this study to acknowledge that DOM objects are not neccessarily d-linked. What is interesting and very important is that *ALL* the examples that are given as examples of test items are those kinds of examples   where the DOM does not differentiate. In number (15) for example bir manken and bir mankeni are not really different in terms of their interpretation. That is, they are objects that have not been introduced prior to their first mention with DOM. So it is not surprising to observe that you do not find any difference between DOM and no DOM objects. It is not clear why you expect to find a difference and why you hypothesize that you will find a difference. Your finding is actually consistent with those literature that you do not discuss throughly. You discuss different kinds of DOM objects in the literature review and then test another kind of DOM objects. Entire literature review and the hypotheses must be reconsidered carefully.

 

Other, relatively less important concerns:

2-      It is not clear how the examples from Nilsson (2) and the statement cited from Comrie are forward looking examples. More detailed discussion must be provided. I

3-      n the last paragraph on the second page (starting line 96), the authors say “Evidence for the forward-looking discouse function of DOM in Turkish also comes from similar examples in other DOM languages.” How can other languages “provide evidence” for Turkish? Do you mean to say, the fact that other DOM languages have a forward looking discouse function suggest that Turkish may have it too? Other DOM languages are not discussed or described in detail, so we don’t know how relevant their pattern to the pattern in Turkish.

4-      Line 182. Note that the categories that you have (definite, indefinite etc) have semantic reference, whereas “bare” is a morphological category. How do you define morphologically bare objects in relation to their referential properties? Or you may categorize all the object types in terms of their morphosyntax, as those that have bir and those that have an accusative case etc

5-      Line 189. It sounds like you are analyzing bare nouns as indefinites. You need to justfy that they are indefinites. You later refer to them as (pseudo)incorporated nominals (Line 194), which contradicts their indefinite status.

6-      Line 274: You say that unmarked direct objects cannot be observed in non-preverbal positions. This is not true. See Ketrez (2023), that you cite, for the references of the literature that discusses such examples.

7-      Line 297: You say that there are no studies that investigates DOM and modification. This is not true either. See Ketrez (2023), that you cite, for examples, references (e.g., Aydemir 2004, Öztürk 2005) and discussion of modification of direct objects by adjectives and relative clauses.

8-      Line 303: “Optional” accusative marking is not the right way to describe the accusative marking in Turkish because there is a significant difference between the case-marked and non-case marked counterparts, as you discuss and the other literature suggests.

9-      Line 306: the statement “variation in case marking is attested only in preverbal position” is not a true. Please see my comment above.

10-   Line 307: “there is no research on modification” is not true, see my comment above.     

11-   How do you distinquish subjects and objects in the bare morphological form?

12-   The semantic category “animacy” makes a distinction between animate and inanimate, it does not distinquish human and non-human animals. You make a distinction between human and inanimate, and exclude the non-human animals (Line 642). Please justify your categorization, which is not the ANIMACY categorization in the literature you cite.

 

Motivation for the study

1-      Lines 124 onwards. Earlier findings suggest that forward-looking function of DOM is not observed in Turkish and you would like to retest the hypothesis. What is the reason behind this motivation? What was wrong with the earlier studies? Why do you think it is worth repeating the test? Please justify your motivation (state, clearly, for example why you think you will find something different, in what ways your research design is superior).

2-      Özge et al and Özge and von Heusinger must be presented and discussed in more detail *with examples from their test items.*

3-      Line 127. What does “broader” empirical approach mean? Analysis of a larger corpora? More participants? More test items? Why do you think this was a problem that needs to be fixed?

4-      Line 196: You say that you will focus on accusative marking with indefinite noun phrases as they show the largest variation. What does “larger variation” mean?

Methodology and analysis

1.       The test items are not presented clearly and they are not provided so the study cannot be replicated in its current form. All test items, control items and fillers must be provided for evaluation either in the manuscript or they must be made available at a data deposit site. So the experiments cannot be evaluated without access to the entire data.

2.       The corpus study excludes an important amount of data and analyzes only a small slice of accusative marked indefinites (non-human animate nouns, abstract nouns, collective nouns are excluded). This is too much manipulation and what you have in the end is not a realistic set of data. If the majority of the accusative marked nouns are abstract nouns in the corpus, then you should be studying the abstract nouns—if you would like to get a realistic behavior pattern of the accusative marked nouns—you should not be excluding them. Apparently, this is the nature of the data. In the experimental settings, participants have access to all these examples that are available in the language when they make grammaticality judgments or when they produce language. So in terms of language processing, the data that you have in your corpus, and the assumed data available to the speakers are not the same kind of data. It is not surprising to find a difference between the corpus pattern and the pattern in the experiments in this sense. So the method of the corpus study must be justified in general.  

3.       Table 1 and Table 2 need illustration. It is not clear what these numbers refer to. The difference between DOM and no DOM does not look significant in Table 1. No statistical analysis was reported, but they are discussed as “different”.

Organization

1-      Section 4.1. is out of place. Previous Studies belongs to the literature review part, not to the section where the present study and findings are presented. This section also repeats what has already been stated in earlier sections.

2-      A totally new study is introduced in the Discussion section. Discussion sections typically include the discussion of the studies that have already been reported in the results section. They do not introduce or present a new study. So the entire presentation of the study on the indefinite article bir should be moved to the results section and must be presented as a follow study (study number 4, following the other 3 studies) and then it should be discussed as such.

3-      The discussion section should present a discussion of the findings in relation to the hypothesis and the literature that was cited. In its current form, it has a summary of the literature, repetition of the hypothesis along with major findings and a totally new study, which does not belong to the discussion section, as stated above. So there is no content that can be considered a general discussion in the general discussion section.

4-      The conclusion section only summarizes the findings which have already been stated a number of times in the manuscript. It should have a conclusion regarding the “implications” of the results (rather than their summary) and include directions for future research. 

5-      The manuscript has a lot of repetitions. Some examples: The hypothesis is stated in (3) and then in (14). If it will not be revised or changed, it does not need to be repeated. You may refer to the previous number. The previous empirical studies Özge et al (2016) and Özge and von Heusinger (2020) discussed on page 3 and then in the same exact words in the Empirical studies section (4.1). Nilsson examples and the Comrie quote are repeated a number of times with exact same wordings. Lines 96-102 and Lines 426-430 are the same, please avoid such repetitions. Example 2 and Example 11 and their discussions are repeated as well.

These need attention:

1-      Line 265, example (6) is not explained. Without context it is difficult to understand these examples.

2-      Line 456. Evidence was presented from Spanish and Romanian??Evidence from Spanish is not really presented or discussed. There is only one sentence about it (Line 426).

3-      Özge et al study (Line 533 onwards.) The stimula is presented but the task was not described clearly. An example from a response would help to illustrate the task. Why did they use ditransitive verbs? And what was problematic about it?

4-      Line 554 why was the study limited?

5-      Please give examples for annotated anaphoric expressions. The description of the study says that the responses were annotated for the following discourse functions—but then they were examined for backward and forward functions of DOM. Shouldn’t they be annotated in both directions for such an analysis? There are inconsistencies in the description of the task.

6-      Line 562: The study assumed that… Do you mean that the study “suggests..”?

7-      Line 579: What does CQP stand for?

8-      Line 640 onwards, you list your variables and you do not include FORM and FUNCTION here. You mention them later, and it sounds like they are your variables, too. So the entire presentation of variables is confusing.

9-      When you say there is a difference, or a relationship is significant or not significant, you should report relevant statistics. There is no statistics reported regarding the data in Table 1 (Line 697).

10-   What does “remarkable effect of DOM” mean? It is also not clear what the statistics reported after Table 2 suggests?

11-   Unmodified and nonmodified terms are used interchangibly. Please be consistent, unless you mean different things.

12-   Line 929: What is item 36?

13-   The examples need to be reviewed and Turkish characters must be displayed properly. Their gloss must be checked as well (in Example 30, for example, the Turkish sentence as an accusative case, but it was glossed as zero).  

Comments on the Quality of English Language

If accepted for publication, the manuscript should be edited in terms of language use. 

Author Response

First of all, we would like to thank all the reviewers for their very constructive and helpful comments. We appreciate the comments and the many very helpful suggestions. We have extensively revised the manuscript to address the comments and added some additional information.

 

Following the reviews, we have reorganized the paper, revised many sections, and added new sections. In particular, we added more motivation for the hypothesis, we expanded the discussion of the parameters for DOM in Turkish, and we reorganized our studies, as we created a new subsection with the first corpus study, which was discussed only in the discussion section in the first version.

Review 1

Thank you for this very detailed and critical review. We very much appreciate that you took so much time to carefully comment our manuscript. We learnt very much from your review and we have tried to address your concerns in the paper. In particular, we have expanded the motivation for our study, we have reorganized the paper in many aspects, and we have also tried to make it clear that we are not talking about all cases of DOM, but about a certain subset that may be quite frequent.

 

Please see the attached PDF for our detailed responses to your helpful comments

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This is a relevant paper, touching on an issue that has seen a strong revival recently, namely the topicality potential of DOM. The data and the methods discussed by the authors are very useful and contribute to scholarship. However, the content needs further clarification at various points. We have included our comments directly on the paper so that the authors can follow them more easily. 

Comments for author File: Comments.pdf

Comments on the Quality of English Language

We have detected only very minor points, as indicated in the comments. These can be very easily fixed. 

Author Response

First of all, we would like to thank all the reviewers for their very constructive and helpful comments. We appreciate the comments and the many very helpful suggestions. We have extensively revised the manuscript to address the comments and added some additional information.

 

Following the reviews, we have reorganized the paper, revised many sections, and added new sections. In particular, we added more motivation for the hypothesis, we expanded the discussion of the parameters for DOM in Turkish, and we reorganized our studies, as we created a new subsection with the first corpus study, which was discussed only in the discussion section in the first version.

Review 2

Thank you very much for your comments in the manuscript, which were very helpful. We corrected the errors in the examples and the alignment (there might be still some mis-alignment due to the conversion from Word into PDF)

Please find our reaction in the attached PDF

 

In the following we try to address and answer your comments and concerns.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

It is well-known, since Enç (1991), the accusative marking in Turkish involves specificity, which can be thought as reference to a subset of the context set, i.e. as partitivity. The author(s) of the paper ask whether ACC in Turkish also has forward-looking discourse function, thanks to which a case-marked object, they claim,  would be more likely to be referred back via an anaphoric expression (referential persistence). They find that the results they got from two experiments do not support this hypothesis.

  The main problem I see with the paper is that the authors seem to rely on an intuitive understanding of "forward-looking discourse function", which is mainly based on Comrie's two claims:
  Part 1: "the absence of the accusative suffix advises the hearer not to bother about identifying the referent" Part 2: "the presence of the accusative suffix advises him that the referent of the noun phrase ... will be of relevance to the ensuing discourse."   With this characterization, the issue becomes one of contrast. this is the spirit of their hypothesis in (3) "Indefinite human direct objects with DOM in Turkish show more forward-looking discourse potential than those without DOM"   Their exposition is mostly about the consequences of having ACC on the noun phrase but it is crucial to their story that Part 1 also holds, i.e. the lack of ACC implies less forward-looking discourse potential, understood in this case as reference via anaphoric elements.I didnt see any discussion in the paper as to why Part 1 should hold. It is obvious that non-case marked objects are routinely referred back via pronominal elements.    Ali sokakta bir çocuk gördü ve ona uzun uzun sarıldı.   So I see two main problems with the paper. First, the theory underlying the hypothesis is not explicit enough. What is a forward-looking discourse function? How do we formally express this?  Second, in generating their hypothesis, they make an assumption about non-case marked objects that they do not defend. But it seems to me that it is necessary that they defend something like Part 1 of Comrie.   Small issues:   on page 19, I couldnt understand what (36) is referring to. on page 19. (30) is quite bad with the accusative marker on tez danışmanı. on page 20, why is there a null morpheme on the advisor?   Comments on the Quality of English Language

no comment

Author Response

First of all, we would like to thank all the reviewers for their very constructive and helpful comments. We appreciate the comments and the many very helpful suggestions. We have extensively revised the manuscript to address the comments and added some additional information.

 

Following the reviews, we have reorganized the paper, revised many sections, and added new sections. In particular, we added more motivation for the hypothesis, we expanded the discussion of the parameters for DOM in Turkish, and we reorganized our studies, as we created a new subsection with the first corpus study, which was discussed only in the discussion section in the first version.

Review 3

Thank you for your comments on how to improve our manuscript. We took your constructive comments into consideration and made corrections and/or highlighted your suggestions in the manuscript. Below we will refer to your points step by step.

Please find our responses in the attached PDF

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments and Suggestions for Authors

The aim of this paper is to explore whether differential object marking (DOM) in Turkish has a “forward-looking” function in discourse. The paper reports the results of several studies using different methodologies with the conclusion that DOM, if at all, only has this function in a very restricted way.

This is a well-written paper that motivates the type of studies it reports on very well. The methods chosen, namely corpus research, forced choice tasks and paragraph continuation tasks, make sense in the context of studying the forward-looking potential of DOM and this empirical approach also serves to validate claims in the literature on Turkish DOM going back to the 1980s (which the paper appears to falsify).

There are only a few minor points I would like to flag. First, when discussing information structure (first on page 2, later on pages 7, 10), topicality is frequently associated with the grammatical function of subject. While this is overall probably valid, I would have loved to see more discussion of information structural properties associated with other grammatical functions in Turkish (independently of DOM). What is generally regarded as topical in Turkish and what evidence is there for the topicality of particular grammatical functions? Outlining the role of information structure in Turkish more generally would provide some additional context for the main discussion in the paper.

Second, regarding the study design discussed in Section 4.2.1, the author(s) describe(s) how they put together their “balanced number of data points”. In this regard, I was wondering whether it is interesting to check the ratio of ACC-marked occurrences of a type to the zero-marked occurrences of the same type. So, if adam appears five times with accusative, how frequently did it appear without? I suspect this ratio to differ across nouns and it could somehow reflect a tendency in how “typical” it is to mark a given noun with accusative (this point is relevant for the discussion of animacy on the bottom of page 21). With respect to the critical items in the forced-choice task (page 16), I acknowledge the large number of variables that are relevant for items like these and the resulting difficulty. It would be very interesting to know more about possible baselines of how likely it is that a subject is the most preferred function to be continued in discourse, or how unusual it is to have overt subjects at all in examples like (22) and (24) given that Turkish can have unpronounced subjects. In addition, certain continuations strike me as less natural for reasons of how the narrative is structured: in (23)/(24), I find it odd for either the apartment manager or the teacher to be smoking a cigarette in front of the building given how the previous discourse is structured. This probably plays no role for the overall results but it might be interesting for the author(s) to check item-by-item results as well.

Finally, there are few minor issues. On pages 2 and 9, when discussing (2) and (11), it is not clarified which (2a) or (2b) (and (11a)/(11b)) the phrase “this sentence” refers to. It becomes clear from the context that it should be the one including overt accusative, but this should be made explicit. On pages 14 and 15, for Tables 1 and 2, it could help to highlight right away what the numbers in the table add up to. For example, below the totals in the right most column in Table 1, it might be useful to add a row showing “186 / 276” in the rightmost column to indicate that the two totals add up to the numbers discussed before Table 1. For clarity, it could also be helpful to make even more explicit that the the 138 items in each row are distinct items, such that for the column “human”, for example, 62/85 and 54/85 do not refer to different fractions of the same 85 items. On page 16, in the second paragraph, I believe the references to (28) and (23) are wrong.

All in all, I believe this paper is a useful contribution to research on DOM in Turkish and provides a good guide on the advantages and difficulties concerning certain methodologies more generally.

Author Response

First of all, we would like to thank all the reviewers for their very constructive and helpful comments. We appreciate the comments and the many very helpful suggestions. We have extensively revised the manuscript to address the comments and added some additional information.

 

Following the reviews, we have reorganized the paper, revised many sections, and added new sections. In particular, we added more motivation for the hypothesis, we expanded the discussion of the parameters for DOM in Turkish, and we reorganized our studies, as we created a new subsection with the first corpus study, which was discussed only in the discussion section in the first version.

Review 4

Thank you for your comments on how to improve our manuscript. We took your constructive comments into consideration and made corrections and/or highlighted your suggestions in the manuscript. Below we will refer to your points step by step.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The revisione made to the paper render the work much clearer, and the manuscript is much easier to read and understand. However, we still find it that the conclusions must be made more explicit right from the introduction. This part could further be improved, a task which nevertheless should not take extensive time and effort. 
Further comments on the organization of the paper, grammar and additional clarification pints have been added to the paper itself. 

Comments for author File: Comments.pdf

Author Response

Thank you very much for the comment and the many helpful remarks in the text. We have revised the text, incorporated your comments and your helpful suggestin,  and also made the argument of the paper clearer from the start. Following your suggestion we made the conclusion clearer at the beginning. Thus, we have revised and significantly expanded an entire paragraph in the introduction.

Thank you once again

Reviewer 3 Report

Comments and Suggestions for Authors

The issues that have been raised are now addressed

Author Response

Thank you very much for your support. Following the suggestions of the other reviewers we have slightly revised our paper and very much hope that it is now even more accessible.

Thank you again for your support.

Reviewer 4 Report

Comments and Suggestions for Authors

I'm happy for this paper to be accepted.

Author Response

Thank you for your support. Following the other reviews we further revised the paper and we think that is now in a very good shape.

Thank you again

Back to TopTop