Re: (OPE-L) Re: Babbage and Marx

From: Paul Cockshott (clyder@GN.APC.ORG)
Date: Wed Mar 31 2004 - 17:21:12 EST

"Gerald A. Levy" wrote:
> Hi Paul C.
> > Depends on the particular product.
> > I illustrate my argument
> > initially with the Roman Samian ware industry which introduced
> > industrialised mass production of consumer goods thanks
> > to adopting moulding technology for ceramics rather than
> > turning them on a wheel.
> > In that case the measure of productivity would be
> > plates produced per worker.
> What often happens with technological advances in the means
> of production is that the material form (and hence use-value) of
> the commodity is altered.  I.e. it often times happens that with
> advances in constant capital that allow for commodities to be
> produced at a lower per unit cost, the material character of
> the commodity is also transformed.  In the context of the example
> you give,  is it only a matter of increased productivity in the
> production of the plates or have the plates themselves been
> materially changed?

What you see when you look at the history of copying technologies
of the sort that Babbage describes in his chapter On Copying
is that they encourage a profusion of elaborate decoration.

Samian ware is almost victorian in its enthusiastic use of
relief, and in this way clearly quite different from previous
pottery. This change in form is objectively measurable in
terms of the product having a higher entropy or information
content. The significant thing about copying technologies,
is that they allow the information content of the product
to be raised by applying parallel transfer of information
to the product. This parallelism is not emphasised by Babbage,
though he clearly understood the principles of parallelism
as is shown by his articles on notation for the description
of mechanical operation, and for that matter in his difference
engine which harnessed parallelism at a very high level.

Babbage could not explicitly apply the concept of information
to production, since, for all his precience, he did not
anticipate Shannon. But he understood that with copying technologies
it was worth lavishing utmost care on the master, from
which copies were made. He estimates that the master may
cost as much as a thousand times what the copies cost.

Although it is evident that Marx borrows very extensively
from Babbage in his analysis of modern industry, he fails to
pay much attention to what Babbage says says about copying
being the key to the reduction of labour content in much
industrial production.

 E.g. is there an increase or decrease in
> embellishment?  Is the shape the same?  Are the non-labor
> circulating inputs the same?  The reason this has significance in
> measuring the extent of the productivity increase is two-fold:
> 1) a change in the final material form of the commodity could
> _itself_ alter the the labor hour requirements to produce the
> commodity.
> 2) if there is a change in the non-labor circulating capital inputs
> this, as well, could alter the total labor hours (or fractions thereof)
> required to produce a unit of output.

A Samian ware bowl involved the production of moulds whose cost
would be considerable, as Babbage argued, but even taking this
into account such copying technologies produce a big net rise
in productivity.

> > The modern world is dominated by copying technology.
> > The improvement in the productivity of the semiconductor
> > industry which has allowed continual exponential growth
> > in production at a rate of around 40% per annum is all
> > based on the perfection of copying technology. In this
> > case the productivity would be measured in terms of
> > transistors produced per labour hour.
> The semiconductor industry is a good example of how
> alongside increases in labor productivity there is also often a
> increased quality of the commodity product.  That increased
> quality can be seen in increases in processor speed.  The 'new
> generation' transistors  could also be seen as a _different_
> commodity (having an altered use-value) being produced.

I deliberately qualified my measure as being in terms of
transistor output not processor output, since the change in
processor design makes them difficult to compare. Functionally
the transistors in successive generations are the same, but
you are right to say that their speed of operation rises, this
is the Win Win nature of MOS technology that Mead and Conway
analysed back in the late 70s and which still applies today
after 25 years of development. Smaller feature sizes produce
a squared rise in productivity of transistors, and at the
same time one gets a linear increase in speed, so if one
defines the ultimate product not to be transistors, but
logical decisions made, productivity rises as the third
power of improvements in the scale at which copies can
be made. It is a believe the first technology whose productivity
development was driven by a cube law. Printing was driven
by a square law, cotton production was linear, most other
technologies were sub-linear.

This archive was generated by hypermail 2.1.5 : Fri Apr 02 2004 - 00:00:02 EST