Jody Hey                  Evolutionary Genetics

  Professor    -     Department of Genetics     -   Rutgers University

By Jody Hey

(ICAPB, Univ. Edinburgh & Dept. Genetics, Rutgers University).

Introduction and Methods

This is a summary of the results of a small experiment conducted by email. The purpose of the experiment was to explore the relationship between apparent clarity of distinction among portions of a graphic and the words used to describe the graphic. The research was motivated by a wish to understand modes of references within scientific contexts (i.e. the nature of the real world and discourse about it) that do not harbor strong distinctions.

While on sabbatical - clearly having way too much time on my hands, some would say - I have been thinking about vague words and vague objects that come up in biology.  However the literature on the relationship among such things is tricky.  On the one hand, there is a body of literature on vagueness as an attribute of language (e.g. Williamson 1994).  But on the other hand, there does not seem to have been much done on how our language reacts to vague objects - on the ways that we describe nature when it does not come as seemingly distinct entities.  Quine does have a bit of discussion on this, but not much (Quine, 1960).  Nor have I found a literature in psychology on the topic.  Certainly perception has been studied (edge recognition etc.), but my questions are more concerned with our descriptions of cases when edges are not to be found.

I figured one way to get a hook into the issues would be to conduct an experiment in which people provided descriptions of pictures that varied in the discreteness of their parts. With a sample of descriptions from different pictures, I could compare the words people had use and see what came out of it. The experiment was not strongly hypothesis driven (though I was surprised by the outcome) but was rather undertaken as an exploration.

The graphics were constructed with a computer program. First I drew this graphic.



Then I used the computer program to blur the picture, which resulted in the second graphic.


 
 

I emailed these pictures to a bunch of folks that happened to be on my list of addresses for one reason or another (nearly all are scientists). Each person that received the email instructions also received one attached file, a jpeg of one graphic or the other (never both). A total of 126 mailings went out for each picture (on Fri Jan 15, 1999), with random assignments of graphics to persons.

The complete email text is here , and the primary instructions were as follows:

Please write down in words a short description of the graphic I have

sent you. Your description should be sufficient to give a reader who

has not seen the picture a fairly good mental image of it. Be terse,

but use complete sentences.

----------------------------

These instructions were designed with the aim of placing the respondent in the particular, but familiar, context of explaining something to someone who cannot see what is being explained.

A total of  86 responses were received, including 41 for the non-blurry picture (sometimes denoted hereafter as N) and 45 for the blurry picture (sometimes denoted hereafter as B).

Pretty much without exception each response included a discrete block of text that was clearly intended as a response to the request. Thus it was a simple matter to remove unrelated text that was sometimes included, and to count the words in the response. Within the responses, only a little bit of editing was needed: beginnings like 'here goes' or references to my sending the picture were removed; '&' was changed to 'and'; '%' to percent; and spellcheck was run.  A few changes were made for easier word counting. Numbers like '1.5'  where changed to 1 point 5. Similarly, expressions like '5x7' where changed to '5 by 7'.  Apostrophes were removed.

Some responses included more than one description and these were taken all together as one response. Often we endeavor multiple descriptions when explaining something, so I saw no reason to tinker with responses that had multiple parts to them.

ã 1999 Jody Hey