Re: Info Mapping : What is a Chunk ?

Subject: Re: Info Mapping : What is a Chunk ?
From: Brad Connatser <cwrites -at- USIT -dot- NET>
Date: Thu, 5 Sep 1996 12:59:46 GMT

Well, I'm in the fray again.

In article <322EFC91 -dot- 2B90 -at- wanadoo -dot- fr>, Frederic Wronecki
<frederic -dot- wronecki -at- wanadoo -dot- fr> wrote:

> A chunk is any physical or abstract stimulus that can be handled as a
> unique mental unit in the memory processing system.

A chunk is not a stimulus. Stimuli are encoded into abstractions similar
to symbols in language (see Smith's Ancestral Voices for a comprehensive
discription of this encoding process).

> Miller's idea is that, at any time, no more than seven stimuli can be
> simultaneously handled.

> But the notion of "chunk" is extremely variable :

> 1 - It may be a visual (text or picture) stimulus, as well as a sound, a
> smell, etc. Miller's principle is then pure common sense : to be
> "distracted" by a nasty smell or a loud noise simply means that your
> global chunk processing capability is partly devoted to the distracting
> stimulus, thus reduced for "useful" stimuli.

I'm not sure what you mean here, but it is true that any code entering
short-term memory takes up space and reduces the amount of space for
information coming in from the automatic reading process (for instance,
when we start day dreaming while we are "reading" and we "read" a whole
page without remembering what we read). Day dreaming is a universal
experience that is, I think, an allegory for your point in no. 1.

> 2 - The "scale" of a chunk is constantly changing : you can observe
> different paintings on a museum wall (each one is a chunk), then focus on
> one of them to examine the characters, the furniture or the landscape,
> thus "forgetting" the other paintings. This splitting process (or its
> reverse : grouping) can of course be "cascaded".

Each painting is not a chunk. Abstractions (call them symbols) can be
chunked, but not stimuli from the optic system.

> 3 - Since it's related to mental processing, a chunk is a different thing
> according to the subject's skill or experience : a professional musician
> will see a set of chords on a score as a single unit, while a beginner
> will have to read each note. A chess master sees the positions resulting
> from the different possible moves as single units, while a beginner will
> have to focus on each position to decompose it, etc.

That's because of the interaction between long- and short-term memory.
Here's an example of how they interact: You're at the closing for the
house you just bought. You have to sign and date a dozen form. But each
time you must date a form, you have to look at your watch for the date.
Why? Because the question "what is the date?" is in short term memory
(synthesized there via chunking from long-term memory and encoded stimuli
from the senses). Once the question is answered, the resulting abstraction
(today's date is ____) is not moved to long-term memory. Instead, the
information is used to activate motors (so that you can write the correct
date). See Pinker's The Language Instinct and anything by Frank Smith (not
STC's honored Journal Editor) for some great examples of this.

> 4 - The "recoding" process helps us make a single chunk from a complex
> set of information by generating a mental summary of it.

> When it comes to writing, all these topics apply, for instance :

> - Like a list, a flowchart should have 7 +/- 2 visible items. If there
> must be more, make some visual groupings.
> - A sentence can be a chunk in some cases, while it's just a part of a
> chunk in others.
> - Info Mapping's "labeling principle" means that you must drive the
> reader's recoding, instead of letting him generate his own label (which
> may be false).
> - An experienced reader will handle each step of a procedure by its
> title, while a beginner will have to read all the paragraph.

Please go back to the "what is the date?" example. Now, here's my argument
for why "chunking"--a psycholinguistic phenomenon--does not neatly serve
as a model for composing technical information: When I'm reading
instructions for assembling a lawnmower, the number of steps it takes to
assemble it is not limited by "seven plus or minus 2." To use Miller's
magic number seven to set limits on the number of steps in a procedure is
arbitrary. Like looking at your watch to answer the question "what is the
date," after the reader of the assembly instructions complete's a step, he
or she can forget it. The number of steps depends upon the complexity of
the procedure, and any conveyance of a procedure should flow from the
reader's need to answer the question "what do I do next?"

I'm not saying that info mapping doesn't work. I'm saying that basing the
labeling principle of infom mapping upon the chunking principle of
psycholinguistics is specious.


Brad Connatser
Concurrent Communications
cwrites -at- usit -dot- net

TECHWR-L List Information
To send a message about technical communication to 2500+ list readers,
E-mail to TECHWR-L -at- LISTSERV -dot- OKSTATE -dot- EDU -dot- Send administrative commands
ALL other questions or problems concerning the list
should go to the listowner, Eric Ray, at ejray -at- ionet -dot- net -dot-

Previous by Author: Re: Information Mapping
Next by Author: Re: Widow "rules"
Previous by Thread: Re: TECHWR-L Digest - 3 Sep 1996 to 4 Sep 1996
Next by Thread: Info Mapping : What is a Chunk ?

What this post helpful? Share it with friends and colleagues:

Sponsored Ads