[Humanist] 26.631 collocation, span and theories of memory?

Humanist Discussion Group willard.mccarty at mccarty.org.uk
Sat Dec 29 11:14:30 CET 2012

                 Humanist Discussion Group, Vol. 26, No. 631.
            Department of Digital Humanities, King's College London
                Submit to: humanist at lists.digitalhumanities.org

        Date: Sat, 29 Dec 2012 09:18:03 +0000
        From: Willard McCarty <willard.mccarty at mccarty.org.uk>
        Subject: collocation, span and theories of memory

I would very much like to understand the historical relation between the 
idea of "span" in psychology, and so the related distinction between 
short-term and working memories, and the ideas of "span" and 
"collocation" in corpus linguistics. I strongly suspect that the former 
inspired the latter but have not yet found an explicit link. Let me 
sketch briefly what I have found so far.

In the OED psychological span is defined as "Mental extent; the amount 
of information that the mind can be conscious of at a given moment, or 
the number of items it can reproduce after one presentation; esp. const. 
of, as span of apprehension, attention span, consciousness span, etc.". 
The first occurrence given is from an article in Mind in 1887, where it 
is clearly used in the spirit of Hermann Ebbinghaus (who began the 
empirical studies of memorization and recall using nonsense syllables). 
In 1956 George A Miller, working by then in the context of digital 
computing, gave a great boost to the idea in his famous paper, "The 
Magic Number Seven, Plus or Minus Two: Some Limits on our Capacity for 
Processing Information" (Psychological Review 63.2). According to Alan 
Baddeley in "Working Memory: Looking Back and Looking Forward" (Nature 
Reviews: Neuroscience 4, October 2003), Miller, Galanter and Pribram 
invented the closely related term "working memory" in Plans and the 
Structure of Behavior (1960), in which the influence of computing is 
fundamental. Short-term memory is not the same but closely related, in a 
manner that varies according to theory, though for my purposes the 
distinction does not matter. What does matter is the teasing 
correspondence between ideas of memory, often involving context-free 
words or word-like units, and structures familiar from the architecture 
of digital computers of the time (memory registers, shifting and 
storage). Kurt Danziger notes in Marking the Mind (p. 176) that long- 
and short-term memory are the only two core-concepts having to do with 
memory that number among the 58 in psychology turned up by a 
content-analysis study textbooks of the 1990s. In the latter half of the 
20th Century there are many publications in the psychology of learning 
and intelligence-testing, some in journals of education, that deal with 
attention- and recall-span measured, again in a Ebbinghausian way, in 
terms of contiguous words remembered.

In the glossary to Corpus Concordance Collocation (1991), John Sinclair 
defines "span" as follows:

> ...the measurement, in words, of the co-text [the words on either
> side] of a word selected for study. A span of -4, +4 means that four
> words on either side of the node [selected] word will be taken to be
> its relevant verbal environment.

In other words, span is the measure within which collocations normally 
occur, +4, -4 for modern English, though Sinclair elsewhere notes that 
it is wise to look more widely. Unfortunately the OED does not record 
"span" in the linguistic sense. In 1957, a year after Miller sprang his 
magic number on the world, J. R. Firth proclaimed in "A Synopsis of 
Linguistic Theory, 1930-1955" , "You shall know a word by the company it 
keeps!" (reapplying the old Latin injunction, "noscitur e sociis"). He 
went on to note that "collocation is first suggested as a technical 
term" in his previously published essay, "Modes of Meaning" (pp. 124-7), 
also published in Papers in Linguistics 1934-1951 (Oxford, 1957).

From the invention of the keyword-in-context concordance (ca 1959, 
three years after Miller's paper, two after Firth's) we have first in 
print-out, then on screen something like a display of a short-term 
memory implying manipulation. From the release of the first truly interactive 
concordancer (TACT, in 1989) we have something like an implementation 
of working memory. I know that I am stretching things here, but not beyond 
the breaking point?

There is a larger story to be told from before the digitization of mind 
by inspiration from physical computing machines -- the stream flows both 
ways -- but meanwhile I would very much appreciate pointers to anything 
that shows explicit transfer of ideas from memory studies to literary 
and linguistic studies.

Many thanks in advance. And comments?

Willard McCarty, FRAI / Professor of Humanities Computing & Director of
the Doctoral Programme, Department of Digital Humanities, King's College
London; Professor, School of Computing, Engineering and Mathematics,
University of Western Sydney; Editor, Interdisciplinary Science Reviews
(www.isr-journal.org); Editor, Humanist
(www.digitalhumanities.org/humanist/); www.mccarty.org.uk/

More information about the Humanist mailing list