What’s the psychological phenomenon that we want to explain?

Title: How far ahead do you plan sentences when you write?

When preparing to produce a short sentence, how much of that sentence is prepared in your mind before you start to speak (or write)?

For a simple sentence like “The cup and the shoe are above the horse” it would be possible to prepare the whole sentence in your mind before you start to speak. There is evidence against that happening [references for where this evidence comes from, unpack one study]. Instead it appears that advance planning has “phrasal scope”. This means that before starting to speak you have to plan the syntax for the noun phrase that is the subject of the verb. In this case that phrase is “The cup and the shoe”.

Possible explanations…

There are two theories about how this might be achieved.

The Syntax First theory is based on the idea that we have an abstract representation of syntax that is separate in our minds from the actual words that we use. Applying this idea to sentence production, the theory is that you retrieve a syntactic “frame” first - an abstract syntactic structure with no content - then slot lexemes into this [add references]. So planning the sentence “The cup and the shoe are above the horse” starts with you retrieving “The ___ and the ___”, and then after that deciding what words you want to put in the spaces. Under this account you don’t have to retrieve “shoe” before you start to speak.

The Lexemes First theory is based on the idea that syntax emerges from the words that we use [references]. So you first retrieve the lexical items that you want to use in your sentence (“cup”, “shoe”) then hang these together in a syntactic structure. Under this account you must retrieve “shoe” before you start to speak.

What do we already know?

Existing evidence for deciding which of these theories is true comes from research in which participants see pictures of objects arranged on the screen and are asked to describe them with sentences like those in the example i.e. with the form “The A and the B are above the C”. The question is whether or not “B” is retrieved before the participant starts to speak.

Previous research has measured the time between seeing the pictures and starting to speak (the “response latency”). Longer time means that participants engaged in more planning. Experiments [references] manipulated how easy it was to retrieve the name for the picture in position B. Making retrieval of the name for the B picture harder will increase response latency only if B has to be retrieved before speaking can start. So only if the Lexemes First theory is true. If the Syntax First theory is true, then manipulating the how easy it is to retrieve B won’t have any effect on response latency.

Evidence from previous research [references] suggests that manipulating the retrievability of B _does_ affect response latency. So this points towards the Lexemes First theory.

But…

However, all of this previous research involved spoken responses. This suggests an alternative explanation, that would be true only for speech. In speech there is a push to be fluent. Hesitation during speaking communicates something to the listener. So you can just stop mid-sentence without this affecting how your listener understands what you are saying. This suggests an alternative explanation for the effects the effects of B on planning. It might not be _necessary_ to plan B for fundamental reasons to do with the cognitive processes behind language production in general. It might just be that in speech people plan ahead more so that they don’t have to hesitate mid sentence, or mid noun phrase.

This effect isn’t there in writing, however. You can pause whenever you like when you are writing and this won’t have any effect on how people interpret what you say.

Preposed research

My study will aim to determine whether the effects found in spoken production - that manipulating the retrievability of B in “The A and the B” affects response latency - are a general feature of language production, or are specific to speech. To achieve this I will conduct an experiment similar to [reference] but will require written rather than spoken responses. If the retrievability effect remains then this will be strong evidence that the Lexemes First theory explains a fundamental feature of the cognitive processes that underlie language production (regardless of whether in speech or writing).

Things to note about writing this…

  1. This is an exercise in explaining your research idea to yourself as much as to other people. Don’t write down anything you don’t understand. Explain things carefully, clearly, and in detail. The important thing is that when you’ve finished writing and discussing your rationale you have a clear understanding of your research.
  2. Think in terms of contrasting two (or more) theories rather than just “is this theory true”. This makes things easier.
  3. These are your theories. Although the two theories I’ve identified can be found in the literature, I’ve simplified them and given them my own names. So I’ve taken control of the ideas. You should do the same.
  4. Don’t get lost in detail. Any clear, plausible, and evidenced argument is good. Read enough other literature to create an argument, but then take control.
  5. Explain your ideas slowly and carefully. The text above is about 750 words. I needed all of those to explain my research idea. And that excludes any discussion of existing literature. If you are writing a lot less than this then you are not explaining yourself clearly.
  6. Use examples. It’s impossible to think about this stuff without them.
  7. This is the first draft of the introduction to your research report. If you get this right all that you’ll need to do when you come to write your introduction in full is add in some references and explain some previous studies. You will them get first class grades for the first two of the six assessment criteria (see details on NOW).
  8. This is difficult. You are working in a conceptually complex area. You’re on a steep learning curve and we are asking you to develop your rationale very quickly. The rationale that you end up with won’t be perfect. It just needs to be good enough to drive your research. Your first version will be nowhere near perfect, but we’ll argue with you until it’s good enough.
  9. This is the real thing. What you are doing here is exactly what all (good) researchers go through when they are planning their research.