Regularisation, systematicity and naturalness in a silent gesture learning task

Yasamin Motamedi, Lucie Wolters, Danielle Naegeli, Marieke Schouwstra and Simon Kirby

While there are 6 possible ways to order the main constituents of a clause – subject (S), verb
(V) and object (O) – they are not uniformly distributed across the world’s languages. Instead,
a considerable majority of the world’s languages use either SVO or SOV order.
A wide body of research has investigated the factors that drive our preferences for different
word orders using the silent gesture paradigm, in which hearing participants are asked to
communicate events using only gesture (Goldin-Meadow et al., 2008; Hall et al., 2013; Meir
et al., 2014). One account, posed by Schouwstra & de Swart (2014), asserts the existence of
natural constituent orders based on event semantics. Extensional events, with direct objects
in motion events, are preferentially expressed with SOV order. In contrast, intensional events,
where the existence of the direct object relies on the meaning denoted by the verb (e.g.
creation events such as bake, build and paint), are more frequently produced with SVO order.
Recently, such ordering patterns have been found in two sign languages, Brazilian Sign
Language and Nicaraguan Sign Language (Flaherty et al., 2018; Napoli et al., 2017),
suggesting that natural orders can persist beyond improvisation in a silent gesture task.
However, we have little understanding of how such distinctions persist, particularly given that
most languages tend towards a systematic word order, where the same word order is used
regardless of event semantics.
We present an online learning experiment in which participants were trained on gesture
sequences representing an extensional and an intensional event, in either SVO or SOV order.
Our design applies an artificial language learning paradigm to the manual modality, building
on previous work investigating regularisation behaviour (Culbertson et al., 2012; Ferdinand et
al., 2019). In a training stage, participants repeatedly saw SVO- and SOV-order gesture
sequences in different proportions in 4 conditions, shown in table 1.
Table 1 Proportions of training orders in each condition
In the natural condition, the majority order was conditioned on the event type as described
above, while in the unnatural condition these natural proportions were inverted. In the
remaining conditions, one order acted as the majority order across both event types. In a
testing stage, participants were repeatedly shown target events and asked to select gesture
sequences with either SVO or SOV order. In this way, we test how participants’ selections of
orders in testing reflect the orders they saw in training, asking whether natural order
preferences persist beyond improvisation, or whether they give way to more regular or more
systematic ordering patterns. Our findings suggest that a bias for natural ordering based on
event semantics may continue to exist beyond improvisation, in competition with a bias for
simpler, systematic ordering patterns.

ConditionIntensional eventsExtensional events
Natural70% SVO, 30% SOV30% SVO, 70% SOV
Unnatural30% SVO, 70% SOV70% SVO, 30% SOV
Majority SVO70% SVO, 30% SOV70% SVO, 30% SOV
Majority SOV30% SVO, 70% SOV30% SVO, 70% SOV
Table 1 Proportions of training orders in each condition


Culbertson, J., Smolensky, P., & Legendre, G. (2012). Learning biases predict a word order
universal. Cognition, 122(3), 306–329.
Ferdinand, V., Kirby, S., & Smith, K. (2019). The cognitive roots of regularization in
language. Cognition, 184, 53–68.
Flaherty, M., Schouwstra, M., & Goldin-Meadow, S. (2018). Do we see word order patterns
from silent gesture studies in a new natural language? In C. Cuskley, M. Flaherty, H. Little,
L. McCrohon, A. Ravignani, & T. Verhoef (Eds.), The Evolution of Language: Proceedings of
the 12th International Conference (EVOLANGXII). NCU Press.
Goldin-Meadow, S., So, W. C., Ozyürek, A., & Mylander, C. (2008). The natural order of
events: How speakers of different languages represent events nonverbally. Proceedings of
the National Academy of Sciences of the United States of America, 105(27), 9163–9168.
Hall, M. L., Mayberry, R. I., & Ferreira, V. S. (2013). Cognitive constraints on constituent
order: Evidence from elicited pantomime. Cognition, 129(1), 1–17.
Meir, I., Aronoff, M., Börstell, C., Hwang, S., Ilkbasaran, D., Kastner, I., Lifshitz, A., Basat,
B., Padden, C., & Sandler, W. (2014). The effect of being human and the basis of
grammatical word order: Insights from novel communication systems and young sign
languages. Cognition, 158(April), 1–40.
Napoli, D. J., Spence, R. S., & Quadros, R. M. de. (2017). Influence of predicate sense on
word order in sign languages: Intensional and extensional verbs. Language, 93(3), 641–670.
Schouwstra, M., & de Swart, H. (2014). The semantic origins of word order. Cognition,
131(3), 431–436.