Re: Cloak paper

Tue, 17 Nov 1998 09:20:14 -0500 (EST)

Subject: Re: Cloak paper
Date: Tue, 17 Nov 1998 09:20:14 -0500 (EST)

On Mon, 16 Nov 1998 16:25:55 -0000 Paul Marsden
<> wrote:
> Re. The Cloak Paper. For the record
here is Cloak's definition of a meme in
> Dawkins 1986 (previous post). Any comments?
> "Memes are neural instructions of a special kind. Neural instructions, in
> turn, are a species of neural control system.

How does the word 'instruction' here relate to the use of that term in
Cloak (1975)? ie:

"the term _instruction_ (....) refer[s] to the unobserved enduring
structure or set internal to the animal."

"an instruction is _any_ [Cloak's italics] material structure that
regularly exhibits a characteristic behavior in response to a
characteristic cue."

"in the black-headed gull [for] example, the (....) instruction
might be depicted as: cue, small white object in nest; behavior,
removal of object from nest"

I like the third one but have reservations concerning the first
two. (all in Cloak 1975, p. 163) You could say that the first
two definitions describe the actual instruction itself and the
last is just about how it is 'depicted'. But this is approaching
the memotype-phemotype dichotomy?? (Is it??) I'd prefer

> A neural control system
> is a
> collection of neurons and synapses organised in such a way that, when
> activated by an impulse from a control system at a higher level, it compares
> a present sensory input to a 'perceptual reference-standard' (Powers 1973)

Haven't read Powers. In fact I haven't heard the term perceptual
reference standard before.

> and until or unless the input approximates the standard, sends repeated
> impulses to one or more control systems at a lower level. Control systems
> at the bottom of the hierarchy send impulses to muscle fibres (and receive
> proprioceptive sensory inputs from those muscle fibres).

These are massively complex arrays of neurons, involving hundreds
of thousand of cells. How do we separate the control aspect from
the gross array of firing neurons?

> If the control system hierarchy is adequately defined, therefore,
> contractions of the muscle fibres usually modify some aspect of the carrying
> organism's present environment, or its relation to its environment, in such
> a way that the present sensory input comes to approximate the perceptual
> reference-standard of the initiating (top-level) control system. At that
> point, the latter stops sending impulses and the entire hierarchy is
> deactivated (Powers 1973).

'If'..... Can we 'adequately define' this?

> A neural instruction is a special control system whose activation
> (behaviour) requires not only an impulse from a higher-level system (or
> instruction), but also a specific stimulus or cue from its environment.
> Like any instruction, in other words, a neural instruction has
> cue-conditions as well as S-conditions [Survival conditions]
> The uniquely defining S-condition of a neural control system or instruction
> may be met by the behaviour of one or more genes and/or by learning
> processes such as operant or classical conditioning. THE UNIQUELY DEFINING
> OBSERVATIONAL LEARNING OR MODELLING (Hall 1963, Bandura 1977)[The capacity
> to form perceptual reference-standards from the basis of verbal behaviour,
> allows memes to be acquired via reading/listening.]

> When an
> animal acquires
> the cue-condition and the perceptual reference-standard of a neural
> instruction by observing the action of another animal responding to that cue
> by approximating that standard, that neural instruction is an
> interchangeable replica of the instruction emplacing the observed animal's
> action (cf. Reynolds 1981 209-213).

Is it though? I'd say the _action itself_ is an 'interchangeable
replica' (and if it isn't a good enough functional replica then
natural selection will weed it out). Here Cloak is sitting
reasonably squarely in the Dawkins B definition. Maybe the
interneural instruction (however, defined, and that itself is
a massive problem which the memetics commuinity has failed to
solve) _is_ an interchangeable replica. But this is a serious
enough statement to require proof, or at least substantial
suggestive justification. Otherwise Occam's Razor comes out.

> Since it may become part of a
> population's traditional behavioural repertoire, the neural instruction so
> acquired is a cultural instruction, or MEME." (Cloak 1986: 163-164)

I'd prefer to say that the 'population's traditional behavioural
repertoire' _is_ the meme pool.

However, I do accept that Cloak's concept of cue+behaviour is an
advance on raw context-free behaviourism.


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)