Evolution and Grammaticalization (3)

Salinas17 at aol.com Salinas17 at aol.com
Wed Mar 1 05:03:49 UTC 2006


In a message dated 2/28/06 2:49:19 PM, oesten at ling.su.se writes:
<< I recommend reading Daniel Dennett's discussion in "Consciousness 
explained"... As will be clear from the following quote, it's not only in linguistics 
that the term is problematic.  "The term "epiphenomena" ... is used with 
*entirely* different meanings... although I have pointed this out time and again, 
no one seems to care."

Perhaps Dennett has made the same unappreciated observations about 
inconsistencies with regard to such terms as "cognition" or "language" or "evolution"?  
He certainly has a lot of opportunity to do a lot more pointing out that 
someone might care about.

<<The two meanings are "a nonfunctional property or byproduct" and "an effect 
which by itself has no effects in the physical world whatever". >>

I'd suggest that the multiple meanings reflect different theoretical 
positions.  And that would also be true with terms like "cognition" and such.  The 
trouble is that this word has a history that doesn't reflect either meaning given 
above.

Without the benefit of Östen Dahl's discussions in his book --
"epiphenomenon" simply meant symptom in its original modern usages, which 
were in medicine.  e.g., "Fever is always secondary to some specific disease or 
other of which it is a mere epiphenomenon or symptom." (1876)  The notion 
obviously was that phenomenon itself was the disease and that the symptom -- 
epiphenomenon -- was peripheral.  Here it is a diagnostic term - a manner of 
characterizing evidence.  But a symptom is a valuable thing in diagnostics -- it's 
not useless or meaningless.

In physics, it has been used to mean nothing more than "irrelevance to the 
matters being observed," i.e, data gathered in an experiment but peripheral to 
variables being measured. In this sense, it was an "operational" definition -- 
a way of distinguishing types of raw data -- impacts appearing on an x-ray 
plate that were not on the principal particle's path.

Given these senses, it's hard to see how it applies to the original poster's 
statement --"Grammaticalization should really be decomposed into its 
independently existing component processes. There's no point in granting explanatory 
power to an epiphenomenon."  

In this sense, epiphenomenon seems to be used to indicate an ineffective or 
unnecessary concept -- not data or symptoms -- which are observations.  A 
concept is not really an epiphenomenon.  Occam's Razor would provide better 
phrasing.

Emergence is something different.  In one sense, "emergence" did mean a 
incidental effect, but it's original use in a psychological sense was by George 
Henry Lewes in the late 1800's, who defined the term as "an effect produced by a 
combination of causes, but not capable of being regarded as the sum of their 
individual effects."  This was in reference to a gestalt kind of approach to 
mind and body.  The whole ending up being greater than the sum of its parts.

Emergent in that sense is an excellent word to use if one is talking about 
sodium and chloride, two poisons, coming together to make table salt.  And it is 
probably useful in describing how human language evolved out of whatever 
pre-language traits preceded it. It goes to unpredictability of effects based on a 
prior examination of causes.

Thus, the use of a phrase like "emergent epiphenomenon" -- in these original 
senses -- is admittedly a little difficult to picture.  But no more so than 
much terminology in a cognitive field that reflects many different theoretical 
approaches and accompanying shades of wording.  If one wants to complain about 
multiple meanings, there's plenty to complain about in other approaches as 
well -- including evolutionary psychology.

Regards,
Steve Long



More information about the Funknet mailing list