About the SPOT game task

Information on the SPOT interactive game-task used to collect speech production data.

What is SPOT?

SPOT is an interactive game-task played by two participants using a specially designed game-board and pieces. Participants negotiate a set of moves to bring the game pieces to their goals. This negotiation uses a fixed set of expressions which are designed for the investigation of how speakers disambiguate sentence structures using intonation and rhythm.

How SPOT came about

SPOT developed out of a common research interest shared by Amy Schafer, Shari Speer and Paul Warren. One northern summer (1997 in fact), Paul visited Shari's former lab in Lawrence Kansas, to discuss possible collaborative research. One area we had separately been interested in working on was a better way of obtaining "naturalistic" recordings of prosodic disambiguation.

Discussion of possible tasks led to the development of a board game with fixed expressions for issuing instructions for how pieces move, but where the successful interpretation of the expressions would involve some pretty neat ambiguity resolution.

Paul had always wanted to devise a game-based task for collecting data, largely because it seemed like taking part in experiments really ought to be fun. Amy, then a post-doc with Shari, proved to have the right mental models for devising cunning board layouts to ensure that we obtained the relevant sets of recordings.

Shari and Paul bought identical sets of wooden play blocks from F.A.O. Schwartz in Kansas City, Paul lugged his back on the plane to New Zealand, and SPOT was born.

The kinds of ambiguity being studied

In brief, the kinds of ambiguities that have long been the subject of interest in psycholinguistic literature on parsing, such as prepositional phrase ambiguity (1), object/complement attachment ambiguity (2), filler-gap dependencies (3), and closure ambiguities (4).

I want to change the position of the square with the triangle
In the context of the board game, this utterance means either "use the triangle piece to move the square piece" (high attachment of the PP) or "move the combined square-with-triangle piece" (low PP attachment - the combined piece resembles a house). Since this ambiguity is not resolved by the words within the utterances it is known as a global or standing ambiguity.
a. I am able to confirm the move of the cylinder
b. I am able to confirm the move was the final one
In utterance a. the noun phrase "the move of the cylinder" is the object of the verb "confirm", whereas in b. the verb's object is a complement clause. The psycholinguistic literature has debated the suggestion that the human sentence processing system 'prefers' the object attachment, and that this is revealed in 'garden path' experiences when the sentence continues as in b. Since this ambiguity is resolved within the sentence, it is a temporary or local ambiguity.
a. Which square_i do you want (t_i) to change the position of the triangle?b. Which square_i do you want to change the position of (t_i) this time?
In a. you want the square to change the position of the triangle, whereas in b. you want something (or someone) to change the position of the square. In other words, "which square" is coindexed with different gap sites in the two utterances, as shown by position of the trace (t). This is again a local ambiguity.
a. When that moves the square, it should land in a good spot.
b. When that moves, the square will encounter a ravenous goat.
In a. the first (subordinate) clause finishes at "square", whereas in b. it finishes earlier at "moves". According to one interpretation of the processing of such constructions, a. - the so-called Late Closure of the first clause - is the preferred structure.

Game boards

You can see one of the game boards in the background to this web site. Others can be inspected by clicking on the links below.

Note that there are different boards for the two participants, who we call the Driver and the Slider, since one guides the movement of the game pieces, and the other slides the pieces in response.

Both boards indicate the starting position of the game pieces, and have the same layout of grid squares, but here the similarity ends. The Driver's board also has the end-points for the pieces, since of course the Driver has the job of getting all the pieces home.

The Slider's board doesn't have these end-points, but does have hazards (ravenous goats) and bonuses (cookies) marked. The Slider has to develop tactics to avoid hazards but encounter bonuses, and this - we hope - leads to certain utterance being used.

Board Driver	Board Slider
Board 1 Driver	Board 1 Slider
Board 3 Driver	Board 3 Slider
Board 4 Driver	Board 4 Slider

Dialects of English being studied

So far, recordings have been made of Midwestern American English (at the University of Kansas) and of New Zealand English (at Victoria University of Wellington). Analysis of both is proceeding.

The meaning of 'SPOT'

We're not really sure. We think it came about because one of the game lines goes When that moves, it should land in a good spot. Since the name arose, we have tried various post-hoc reconstructions of a possible acronym, such as "Speech Production Orientation Task", but we're still open to suggestions.

Why the goat?

Best ask Amy. And while you're there, praise her for the artwork. You might like to note that in Wellington at least, participants get to keep any cookies they land on. We didn't give away any goats.

Acknowledgements

This work received support from several sources.

The NZ/USA Cooperative Science Fund supported Paul Warren's first visit to Lawrence Kansas in 1996, when SPOT was born.
The Marsden Fund, administered by the Royal Society of New Zealand, supported Paul Warren's research on intonation (including the SPOT project) through grant VUW604.

School of Linguistics and Applied Language Studies Te Kura Tātari Reo