Re: memetic speciation?

John Wilkins (wilkins@wehi.EDU.AU)
Mon, 23 Jun 1997 12:24:59 +1100

Date: Mon, 23 Jun 1997 12:24:59 +1100
From: John Wilkins <wilkins@wehi.EDU.AU>
Subject: Re: memetic speciation?
In-Reply-To: <>

>From: Bill Benzon <>
>Date: Fri, 20 Jun 1997 12:12:52 +0000
>Subject: Re: memetic speciation? [was what's in a meme]
>If Price asks:
>> a. what sense of genentic do you mean. I assume the general rather
>> than
>> specifically gene based.
>Yes. The term "genetic" is common in historical linguistics, but it
>doesn't imply anything about biology. It's strictly a cultural term.
>> b. I was not suggesting a tree. Sure linguistic species cross
>> pollenate as
>> they evolve.
>I don't in fact know if the term "taxonomy" implies a tree structure,
>but trees are certainly a common classification structure. [...]
>So, where do you put creoles in the classification system? If you look
>at Ruhlen's book you'll find he has a Creole taxon at the top-level of
>his system. That implies that these languages have no parent language
>but the Ur-language of them all, which simply is not true. They have
>very identifiable parents - 2 or 3 or even more of them. But the
>classification system won't let you put such languages somewhere down in
>the structure with links to all of the parents.
>Of course, classification systems of all but the simplest collections of
>objects are going to have difficult cases. The existence of difficult
>cases shouldn't automatically invalidate the classification scheme. But
>in the case of language, we have a methodology which is designed to
>produce a certain kind of result. In that case I'm not sure it is wise
>to allow creoles to be treated as problematic exceptions. It is clear
>that creolization produces genealogies which do not have the form of a
>tree. And thus it seems to me at least possible that some or many of the
>difficulties down in the taxonomic tree exist because there has been
>extensive creolization for which we have no evidence in the historical
>or archeological record.

A short report from the depths of my in-progress PhD:

Taxonomies have used something like a tree structure since Plato, and one
famous variation is called Porphyry's Tree, which is a variety of Hennigian
Comb. Aristotle classified in two ways, one of which involved 'privation',
or dividing into dichotomies of those which do possess a property and those
that don't. Whether all taxonomies must use a tree is moot.

Hennig's classification - cladism - involves a tacit or at least
instrumental assumption that speciation is divergent, and that each event
is logically a dichotomy. Another way to say this which is relevant here is
that descent below the species [language] level is assumed to be reticulate
while above it, it is assumed to be purely divergent. Phyletic information
useful for a cladistic analysis is lost when reticulation occurs. Without
some paleontological information, the One True Phyletic Tree cannot be

Creolisation is a very good example of the sort of problem that arises in
plant evolution and taxonomy. I view taxonomy as classification by descent,
but this is not the standard view by any means. There are taxonomies of
logical methods and taxonomies of minerals and crystals. I just happen to
think that if taxonomy is not identical to classification, then it needs to
have some basis that forces it to exceed mere convenience, and descent
relations strike me as the best choice for "natural kindhood".

IMO the appropriate metaphor for a phyletic classification (which does not
exhaust the set of all classifications by descent) is not a tree, nor even
a bush, but a time-variant net.

John Wilkins
Head of Communication Services
Walter and Eliza Hall Institute

This was distributed via the memetics list associated with the
Journal of Memetics - Evolutionary Models of Information Transmission
For information about the journal and the list (e.g. unsubscribing)