Implicit Social Meanings (Implicit Association Test - IAT and keyword association data) associated with the use of infinitivial or finite (da+present) clausal complements in Serbian and Croatian

This dataset, Implicit Social Meanings Associated with the Use of Infinitives and Finite Clausal Complements in Serbian and Croatian (IAT, Keyword, Regional Association, and Acceptability Judgment Data) v2.0, provides valuable insight into the nuanced interplay between linguistic structure and social identity in the Shtokavian dialect continuum. Derived from an experiment conducted in 2024 at the Universities of Novi Sad and Niš as part of the ClaCoLaSI project, it sheds light on regional variation and implicit social perceptions tied to syntactic choices in Serbian and Croatian.

Key Features of the Dataset:Geographical Focus:

Vojvodina (Northern Serbian Province): High frequency of infinitive usage.Rest of Serbia: Lower frequency of infinitives, higher reliance on finite complements.Croatia: Infinitives dominate, with finite complements being rare.

Sample:

The sample consists of 324 participants, including 116 from Northern Serbia (Vojvodina), 125 from the rest of Serbia, and 80 from Croatia. Additionally, data from three participants from neighboring countries are included, with decisions on their inclusion or exclusion to be made on a case-by-case basis depending on analytical goals. Participants were recruited voluntarily from students at the Universities of Novi Sad (Vojvodina), Niš (South Serbia), and Rijeka (Croatia) using a snowballing method, where students were encouraged to recruit other participants. The gender distribution of the sample includes 248 females, 74 males, and 2 non-binary individuals. The mean age of participants is 27.75 years (SD = 9.79). In terms of education, 183 participants had high school education at the time of the experiment, 108 had college degrees, and 33 held graduate degrees.

Objectives:

To isolate implicit social meanings linked to the use of infinitives and finite clausal complements.To measure regional variation in these meanings across Serbia and Croatia.To explore how linguistic preferences reflect broader social and cultural identities.

Data Collection Procedure:Participants, all native speakers of Serbian or Croatian, completed an online survey. They provided demographic information—age, gender, place of residence, region where they completed elementary education, education level, and self-assessed proficiency in Serbian/Croatian and several foreign languages (English, German, French, Russian, Italian). Following this, each participant was randomly assigned one out of 24 possible target sentences (12 syntactic frames × 2 variants [INF vs. DPC]) to evaluate, ensuring a balanced sampling of different structural configurations.

Target Sentences and Linguistic Variables:

Syntactic Configurations: The 24 target sentences were constructed around 12 syntactic frames, each of which could appear with either an INF or a DPC complement. These 12 frames were themselves organized around 6 different configurations known to allow both complement types, but which vary in markedness and usage frequency:Clausal Subject Position:Sentences in which the INF or DPC functions as the subject of the clause.Complements of a Future Auxiliary:Configurations where an INF or DPC is embedded under a future-tense auxiliary verb.Stacked Complementizers (Cognition/Speech + Modal/Volition):Sentences where a verb of speech or cognition, which obligatorily takes the complementizer “da,” is followed by a modal verb or a volitional verb. In these frames, the embedded verb can appear in either INF or DPC form, resulting in a “stacked” da construction or a da+INF pattern.Embedded under a Modal with Epistemic Reading:Frames where an epistemic modal verb governs an embedded clause, allowing for an INF or DPC complement. Prior research suggests that DPCs are degraded in these contexts.Embedded under a Verb of Attempt (e.g., “try”):Constructions in which a verb meaning “try” selects an INF or DPC complement.Modal with an Abstract Subject:Scenarios in which a modal verb appears with a non-specific, abstract subject, and can be followed by either INF or DPC. Previous work indicates that DPCs may be marked or dispreferred in these contexts.There were two distinct sentence frames within each of these six configurations, totaling 12 frames. Each frame appeared once with an INF and once with a DPC, generating 24 unique sentences. Each participant rated only one of these sentences, ensuring that evaluations were not influenced by repeated exposures.

Rationale for Frame Selection:The selection of these frames was guided by both normative grammatical commentary and existing linguistic research:

Public Discourse and Normative Commentary:Certain configurations involving DPCs—especially those in subject positions, under a future auxiliary, or with “stacked” da complementizers (e.g., after a verb of speech/cognition plus a modal or volitional verb)—are frequently criticized by normative grammarians and public discourse as markers of poor style or substandard language.

Linguistic Research and Markedness:Other configurations (e.g., DPCs under epistemic modals, with abstract subjects, or in attempt constructions) have not been widely discussed in normative literature. However, previous studies (Kovačević et al. 2018; Arsenijević, Kovačević, Milićev, in press) have shown that DPCs are marked and often degraded in these contexts, even though such judgments are not commonly reflected in public commentary.

By including configurations that are both publicly stigmatized and those that are less frequently discussed, the study seeks to determine if the social negativity and decreased acceptability associated with certain uses of DPC match public commentary and whether similar attitudes emerge in more subtle contexts not previously highlighted in prescriptive discourse.

Geographic and Cross-Linguistic Variation:Another key motivation is to explore variation across different speaker communities:

Croatian vs. Serbian Contexts:DPCs are highly marked and relatively rare in Croatian, whereas INF use is more common. By comparing Croatian speakers to Serbian speakers, we can assess whether norms differ based on language identity and national standard.

North vs. South Serbia:Within Serbian-speaking communities, INF forms are more common in the North and much rarer in the South. The dataset thus allows for the examination of regional variation and whether geographic origin influences acceptability judgments and social evaluations.

Experimental Tasks:

Acceptability Judgment:Each participant provided a grammaticality rating of the single assigned sentence, which varied in structure (INF vs. DPC) and syntactic configuration (one of the six configurations described above).

Speaker Evaluation Scales:Participants rated the speaker of the sentence on 10 social attributes (see accompanying “Scales” file), probing perceptions such as intelligence, socioeconomic status, trustworthiness, and cultural refinement.

Cultural and Social Orientation Measures (Palate Cleansers):

Favorite Food Choice: Participants selected from options spanning national vs. international cuisines and upscale vs. downscale establishments. This choice serves as an indirect measure of cultural and class orientation.Open-Ended Adjectives: Participants provided three adjectives describing the speaker, yielding qualitative data on social perceptions.Music Choice: Participants chose from music options representing national/international and high-culture/popular variants, further reflecting cultural leanings.

Regional Attribution:Participants were asked to guess the speaker’s regional origin from a list of 10 major ex-Yugoslav urban centers. This measure checks whether particular syntactic forms are linked to regional stereotypes.

Attention Check:A clearly ungrammatical sentence was presented at the end, and participants rated its acceptability to confirm their attentiveness.

Data Contents:The dataset includes:

Demographic Variables: Age, gender, place of residence, place of elementary schooling, education level, self-assessed language proficiency in Serbian/Croatian and several foreign languages.Linguistic Variables:Sentence type (INF vs. DPC)Syntactic configuration (one of the six defined configurations, specified via frame identifiers)Acceptability ratings of the target sentence and the final ungrammatical control sentence.Social Evaluation Variables:Ratings on the 10 social scales.Food and music choice indicators (coded to reflect national/international and high/low cultural value).Three adjectives describing the speaker (qualitative data).Regional attribution choice (one of 10 cities).

Data Format and Documentation:The dataset is provided in a tabular format (e.g., CSV) suitable for quantitative statistical analysis. Variables are coded and documented in accompanying README files, with codebooks specifying variable names, coding conventions, and instructions for interpretation.

Usage Notes:Researchers in sociolinguistics, syntax, language attitudes, and social psychology may find this dataset valuable. It supports investigations into how grammatical variation is linked to social meanings, cultural orientations, and regional identity perceptions. By encompassing both publicly stigmatized structures and less discussed configurations, the dataset allows for a nuanced exploration of the interplay between syntax, normativity, and social evaluation across different speaker communities in the South Slavic linguistic area.

FILES:

Dataset: Implicit Social Meanings Associated with the Use of Infinitives and Finite Clausal Complements in Serbian and Croatian (IAT, Keyword, Regional Association and Acceptability Judgments Data) v2.0

README file explaining the data structure: README Implicit Social Meanings Associated with the Use of Infinitives and Finite Clausal Complements in Serbian and Croatian (IAT, Keyword, Regional Association and Acceptability Judgments Data) v2.0

Stimuli and code - presenting the original stimuli for the experiment and 

References

Arsenijević, B., Kovačević, P. & Milićev, T. (in press). A Concreteness Effect in Morphosyntactic Variation: Evidence from the variation in the choice of Clausal Complements in Serbian. Psihologija.  

Kovačević, P., Milićev, T., & Đurić-Paunović, I. (2018). The variation in non-finite complements in Serbian: empirical evidence at an intra-speaker level. Annual Review of the Faculty of Philosophy, 43(1).