This is a review of the paper - " Children’s comprehension and production of transitive sentences is sensitive to the causal structure of events" by Melissa E.Kline, Jesse Snedeker and Laura E. Schulz presented in "Proceedings of the Thirty-Second Annual Conference of the Cognitive Science Society"
Introduction
One of the most interesting aspects of human behavior is our ability to interpolate and predict events. We are able to do this because of the reason that most of the events happening around us are Causal in nature (Deterministic View). Adults as well as young children learn about this causal nature of events primarily through 2 methods:-
1) First Person Observation.
2) From the language that they hear
My review will focus on how this paper tries to establish a relationship between language acquisition (i.e. understanding and forming transitive sentences) and the causal structure of events in young children and then I will also try to correlate the results with Michotte's video.
Some Defintions
Now before starting with the review, let us define some basic terminology used:-
1) Causality: It is basically defined as the relationship between two events (say A&B) such that B can be unnderstood as the effect of A.
2) Language Acquisition: "Language acquisition is the process by which humans acquire the capacity to perceive, produce and use words to understand and communicate. This capacity involves the picking up of diverse capacities including syntax, phonetics, and an extensive vocabulary."[2]
3) Transitive sentences: These are sentences wherein the verb answers a 'What' or a 'Whom' question. It has a subject, a predicate and an object. Example: "The cat spilt the milk." It answers the question; The cat spilt what??
Motivation behind the work
If we look at the various transitive sentences we encounter everyday (e.g. Sumit broke the lamp), we tend to find a cause (Sumit) and an effect (broke the lamp) there. Combining this with some of the other factors enlisted next, we get reasons to think that 'yes' adults and children do tend to interpret transitive sentences causally. Some of the main reasons are:-1) If we consider languages other than English (such as Hindi or English), there is a tendency to express verbs of direct external causation with transitives.2) The paper talks of an experiment wherein some adults were asked to guess the meaning of some nonsensical transitive sentence like ‘The glob blebbed the meemor’. Now the adults reported that such sentences were likely to describe events which had properties of physical causation: motion, contact, exertion of force and causation.All these examples suggest that there might be a link between the transitive syntax we use and the causal structure of events.
Previous work and challenges
Some of the previous work done in this field and their results are as follows:-1) The work done by (Leslie & Keeble 1987; Muentener & Carey, 2010) showed that in nonlinguistic tasks, young infants are sensitive to spatiotemporal cues in distinguishing causal and noncausal events. This ability allows them to learn about the world and design effective interventions on causal processes.2) The work of (Harris & Koenig, 2006) showed how children learnt a great deal about the world around them through the narrations of others.
Apart from these studies, there are still some challenging areas about which we don't know a lot and this paper has tried to study some of these fields.1) How do children map causal relationships into language?2) We have very little understanding of whether children selectively produce transitive sentences to explain scenarios where the cause and effect events obey spatiotemporal contiguity. Inversely, we also do not know whether children selectively expect transitive verbs to refer to causal events. This paper tries to answer the second question in detail using 2 sets of experiment. Experiments performed in this field earlier used a very coarse approach wherein the children were given scenes which contrasted on a wide variety of fields. On the other hand, this study uses a minimal pair approach to see whether a particular factor (causality) drives young children's expectation towards transitive sentences or not.
Experiment 1
Aim
The main aim of this experiment was to look at "the production of transitive sentences"
Participants
The participants were preschoolers (n=24, mean age = 3)
Procedure
In total 4 novel actions were setup. All these experiments were initiated by a puppet held by the experimenter. The causal versions of these were:-1) Meeking: Puppet reaches both hands out to touch blinking fiber-optic wand.2) Wugging: Puppet goes down to contact round globe which lights up and spins.3) Gorping: Puppet lands on a toy which squeaks.4) Pilking: Puppet places ball on ramp, which rolls down to hit a donut shape which boings. All the events had both the causal as well as the non causal counterpart, thereby making the total number of cases to be 8. The non-causal version differed from the causal ones by the fact that there was a 10-15 cm gap and a 1 sec pause between the puppet's final position and the toy activating. A sample video of causal wugging and noncausal wugging is given below:-
The experimental process as mentioned in the paper is as follows - "Children were introduced to ‘Sarah’, a puppet who liked to say silly words. Children were prompted to repeat two novel words (zorb, gliffit) to prepare them for the test session. They were then introduced to the second puppet (Joey) in the same manner. Each participant saw one version of each of the four events, with causal and noncausal events alternating. Event version and order of presentation was counterbalanced across children. Children received the same linguistic and event exposure for each trial. Children watched the action (enacted with Sarah as the agent)three times, with the following description: When this happens, that’s called wugging. Let’s watch again, because the round thing is gonna wug. Whoa, it’s wugging! Now Sarah’s gonna have one more turn. [Prompt 1]: Can you tell me what she’s gonna do with that thing?[Prompt 2]:What’s gonna happen? If children didn’t respond to prompts 1-2, they were reminded of the name of the event and prompted again to describe the event: [Prompt 3]: ‘When this happens it’s called wugging. So, can you tell me what she’s gonna do with that thing?’ Finally, the second puppet was introduced and the entire elicitation procedure (Now Joey’s gonna have a turn...) was repeated." [1]
Results
The initial prediction by the paper was that children would produce a transitive generalization more for a causal event than for a non causal one. The observed results were as follows:-1) Children produced transitive sentences on more causal trials than noncausal. To add to that, they readily produced transitive explanations of the causal events but required all the three prompts for the non causal one.2) The number of trials required to produce an explanation for the causal trial was less than the non causal trial.3) 4 year olds were able to make more transitive sentences than the 3 year old ones.
Experiment 2
Aim
The main aim of this experiment was to find out how children preferentially select a scene preferences following a transitive sentence.
Participants
The participants were preschoolers (n=24, mean age = 3)
Procedure
The 4 scenarios used in expt. 1 were coneverted into videos. Two new events - Fooping: Puppet touches the balanced blue wedge which tips over and swings, Zigging: Puppet slides over and makes white 'accordian' rectangle pop out.
The start up of the experiment was mostly the same as that of Expt. 1. However there was one addition in which Sarah showed them a helicopter apparatus, showing that sometimes, Sarah puts her hand on top of the box and makes it go. But sometimes, it just happenson its own, because there’s a battery inside. The children were then made to do the same thing. Then Sarah explained that in the movies to foloow sometimes Sarah would make things happen and sometimes it would happen on their own. In the final experiment, the children were shown the causal as well as the non causal version of the events on the same screen. The order of the trial and the side of the presentation was randomized to ensure no biasness. (See Video). Children were then asked 2 questions:- 1) Can you find the movie where she meeked the round thing?2) Can you find the movie where she didn’t meek the round thing? The order was randomized across trials. In order to check manipulation, children were also asked to identify ‘the movie where they were touching.
Results
1) For the first question, the distribution of the scores were above chance performance i.e. no children chose less than 2 causal scenes in response to the prompt.2) For the second question, the distribution of the scores were below chance, thereby indicating that the first result was not just an outcome of global preference for causal movies.3) Age did not play much of a role here. The experiment thus showed that the transitive sentences were mapped correctly onto the causal choices.
Conclusion
Both of the experiments performed, show that the causal relationships are very important for children's expectations about transitive sentences. Children are sensitive to causality when it comes to deciding the use of new/novel verbs. The work done in this paper also shows how children are able to map different types of causal events into language, thereby also showing how they learn to us language like an adult and also learn about causal relationships from the world around them.
Michotte's Video and it's relationship with the paper
The following representation of Michotte's work was obtained from http://openmap.bbn.com/~kanderso/Michotte/michotte.html
Second demonstration onwards, the video does not seem to have any causal relationship between the blue and the red ball. However, if you ask anyone a transitive question such as what happens to the blue ball in the video, he/she would probably give a causal explanation of the thing. This is becauase (as pointed out in the paper), we have a causal expectation to the transitive sentences thrown at us mostly because that is the way we map events into language. It would be very interesting to perform the Michotte's video experiment on children and see how things workout.
Acknowledgement
1] Kline, M., Snedeker, J., & Schulz, L.E. (2011). Children's comprehension and production of transitive sentences is sensitive to the causal structure of events. Proceedings of the Thirty-Second Annual Conference of the Cognitive Science Society and the Supplemental material on the website for the videos.2] Wikipedia - for certain definitions.3] http://openmap.bbn.com/~kanderso/Michotte/michotte.html for Michotte's explanatory video.4] Cognitive development, culture, and conversation: comments on Harris and Koenig's "truth in testimony: how children learn about science and religion" - Harris & Koenig, 2006 - It is a very interesting read