Why we need a gradient approach to word order

Levshina, Natalia; Namboodiripad, Savithry; Allassonnière-Tang, Marc; Kramer, Mathew Alex; Talamo, Luigi; Verkerk, Annemarie; Wilmoth, Sasha; Rodriguez, Gabriela Garrido; Gupton, Timothy; Kidd, Evan; Liu, Zoey; Naccarato, C.; Nordlinger, Rachel; Panova, Anastasia; Stoynova, Natalia

doi:10.31234/osf.io/yg9bf

Cited by 7 publications

(8 citation statements)

References 199 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Currently, the dominant theories in Tagalog syntax offer rules and principles that provide an explanation for a language’s flexible word order, and presume processing explanations for the attested frequencies or probabilities in the language. These probabilities, illustrating the dominance and variability of word order, support a gradient approach to explaining flexible word order in Tagalog, similar to what has been described in other languages such as Malayalam (Levshina et al, 2021; Namboodiripad, 2017, 2019).…”

Section: Methodssupporting

confidence: 77%

Differential effects of agency, animacy, and syntactic prominence on production and comprehension: Evidence from a verb-initial language.

Bondoc¹,

Schafer²

2022

Canadian Journal of Experimental Psychology / Revue canadienne

View full text Add to dashboard Cite

Effects of animacy and agent prominence in linguistic and cognitive processing are well-established in the literature. However, it is less clear how strongly an agent argument will influence production and comprehension when a sentence also contains another prominent argument. We examine this question with Tagalog, a verb-initial language, which designates a syntactically prominent, subject-like element (the pivot) without demoting the grammatical status of the core agent. We implemented two experiments that investigated the influences of agent and pivot prominence on syntactic linear word order patterns in production and on anticipatory gaze patterns in comprehension. Tagalog’s grammar allowed us to separate the influence of agentivity from animacy by manipulating the animacy of the pivot (animate pivots: agent and benefactive voices; inanimate pivots: patient and instrument voices). The production results contrasted with the comprehension results: agent and pivot prominence both emerged strongly in a fragment-completion production task, but animacy dominated anticipatory gaze patterns in a visual-world comprehension task. The results of these experiments demonstrate variability in production and comprehension outcomes as well as an apparent mismatch between the constraints that shape these two systems, which we attribute to contrasting goals in production versus comprehension and to the organization of information in verb-initial languages. The investigation highlights the value of research on languages with typologically understudied structural properties in revealing mechanisms of the production and comprehension systems.

show abstract

Section: Methodssupporting

confidence: 77%

Differential effects of agency, animacy, and syntactic prominence on production and comprehension: Evidence from a verb-initial language.

Bondoc¹,

Schafer²

2022

Canadian Journal of Experimental Psychology / Revue canadienne

View full text Add to dashboard Cite

show abstract

“…We do find that the prior existence of a conventional lexicon does, in line with previous findings, lead to an overall increase in the preference for SVO-like structures, but that this is itself mediated by the semantics of the event and by participants' existing linguistic knowledge. This interplay of different factors in governing participants' ordering preferences is consistent with the gradient and contextdependent nature of ordering patterns in natural languages (Levshina et al, 2021;Napoli & Sutton-Spence, 2014), and calls for a nuanced approach to the cognitive and contextual processes that influence language structure.…”

Section: Discussionmentioning

confidence: 55%

“…Previous silent gesture research has suggested that a preference for SVO orders can be explained with domain-general accounts, namely a bias to place subjects first regardless of the other constituents, and the robustness to noise of SVO order, particularly for reversible events where subject and object can be confused (Gibson et al, 2013;Hall et al, 2014). Moreover, evidence from both natural languages and lab-based experiments have shown that ordering preferences are sensitive to a wide range of language-internal and external factors, including event semantics and reversibility, as well as animacy, the surrounding linguistic context, and the social-interactive context in which language is used (Christensen et al, 2016;Gibson et al, 2013;Kirton et al, 2021;Levshina et al, 2021;Meir et al, 2014;Napoli & Sutton-Spence, 2014;Schouwstra & de Swart, 2014). Ordering patterns within and across languages will reflect this interplay of different factors.…”

Section: Discussionmentioning

confidence: 99%

The effects of iconicity and conventionalisation on word order preferences

Motamedi¹,

Wolters²,

Schouwstra³

et al. 2021

Preprint

View full text Add to dashboard Cite

Of the 6 possible orderings of the 3 main constituents of language (subject, verb and object), two —- SOV and SVO —- are predominant cross-linguistically. Previous research using the silent gesture paradigm in which hearing participants produce or respond to gestures without speech, has shown that different factors such as reversibility, salience and animacy can affect the preferences for different orders. Here, we test whether participants’ preferences for orders that are conditioned on the semantics of the event change depending on i) the iconicity of individual gestural elements and ii) the prior knowledge of a conventional lexicon. Our findings demonstrate the same preference for semantically-conditioned word order found in previous studies, specifically that SOV and SVO are preferred differentially for different types of events. We do not find that iconicity of individual gestures affects participants’ ordering preferences, however we do find that learning a lexicon leads to a stronger preference for SVO-like orders overall. Finally, we compare our findings from English speakers, using an SVO-dominant language, with data from speakers of an SOV-dominant language, Turkish. We find that, while learning a lexicon leads to an increase in SVO preference for both sets of participants, this effect is mediated by language background and event type, suggesting that an interplay of factors together determine preferences for different ordering patterns. Taken together, our results support a view of word order as a gradient phenomenon responding to multiple biases.

show abstract

“…Our lab observations are consistent with word order data on reversible vs. non-reversible events from existing languages in that respect, but then we should ask the question: if having a consistent word order helps in reliably conveying information about reversible events, why do many existing languages allow for variation? All in all, the data indicates that the way in which word order conventions are formed in real life is probably more complex than we have sketched so far, an observation that is central in Levshina et al (2021), who propose that linguists need to approach word order as a gradient phenomenon rather than labeling languages as belonging to fixed categories. Another option is that the SOV preference for nonreversible events and the SVO preference for reversible events are in fact stronger than the more general bias to use word order completely consistently.…”

Section: Word Order Regularization In Communicative Interactionmentioning

confidence: 84%

Investigating Word Order Emergence: Constraints From Cognition and Communication

2022

View full text Add to dashboard Cite

How do cognitive biases and mechanisms from learning and use interact when a system of language conventions emerges? We investigate this question by focusing on how transitive events are conveyed in silent gesture production and interaction. Silent gesture experiments (in which participants improvise to use gesture but no speech) have been used to investigate cognitive biases that shape utterances produced in the absence of a conventional language system. In this mode of communication, participants do not follow the dominant order of their native language (e.g., Subject-Verb-Object), and instead condition the structure on the semantic properties of the events they are conveying. An important source of variability in structure in silent gesture is the property of reversibility. Reversible events typically have two animate participants whose roles can be reversed (girl kicks boy). Without a syntactic/conventional means of conveying who does what to whom, there is inherent unclarity about the agent and patient roles in the event (by contrast, this is less pressing for non-reversible events like girl kicks ball). In experiment 1 we test a novel, fine-grained analysis of reversibility. Presenting a silent gesture production experiment, we show that the variability in word order depends on two factors (properties of the verb and properties of the direct object) that together determine how reversible an event is. We relate our experimental results to principles from information theory, showing that our data support the “noisy channel” account of constituent order. In experiment 2, we focus on the influence of interaction on word order variability for reversible and non-reversible events. We show that when participants use silent gesture for communicative interaction, they become more consistent in their usage of word order over time, however, this pattern less pronounced for events that are classified as strongly non-reversible. We conclude that full consistency in word order is theoretically a good strategy, but word order use in practice is a more complex phenomenon.

show abstract

Why we need a gradient approach to word order

Cited by 7 publications

References 199 publications

Differential effects of agency, animacy, and syntactic prominence on production and comprehension: Evidence from a verb-initial language.

Differential effects of agency, animacy, and syntactic prominence on production and comprehension: Evidence from a verb-initial language.

The effects of iconicity and conventionalisation on word order preferences

Investigating Word Order Emergence: Constraints From Cognition and Communication

Contact Info

Product

Resources

About