Re: Books on Ashby's Law of Requisite variety?

E. Andres Garcia (egarcia@STERN.NYU.EDU)
Thu, 16 Jan 1997 20:43:16 -0500


This appears to be one more potential misapplication of Ashby's Law of
Requisite Variety, which I'm becoming more and more convinced is not a
"law" at all. At least two problems come to mind:

1) One problem lies in the observer's necessarily highly subjective
perception/interpretation of "variety", both in the system under
investigation and its environment (an arbitrary distinction to begin
with). You could easily end up with a tautology, where if a
system/organism does not survive in an environment full of variety, then
the observer concludes that the system/organism just did not embody enough
variety itself (regardless of the observer's ability to observe the
variety), while if the system/organism does survive, well then it clearly
did have enough variety. Our inability to objectively identify variety is
thus intimately related to our inability to objectively identify
evolutionary selection criteria within a constantly shifting fitness
landscape.

2) Another problem lies in the evolutionary selection mechanisms operating
at different levels of analysis (for instance, at the individual, group,
organizational, societal, and ecosystem levels). These potentially operate
at cross-purposes: for example, in the well-known metaphors of the
"Tragedy of the Commons" or the N-Person Prisoner's Dilemma, short-term
pressures selecting for individual self-interested behavior end up
potentially causing the entire system to collapse and not survive. If
natural selection is thus perceived as a fractal process, then there is no
"law" guaranteeing that an environment with high "variety" (however you
choose to subjectively define it) will proceed to select systems/organisms
also embodying a similar level of variety; what ends up surviving could
very well be relatively quite simple (viruses, for example).

We need to seriously reexamine Ashby's "Law". It was formulated within a
first-order cybernetics paradigm and probably requires serious revision.

Andres Garcia
NYU Stern School of Business
egarcia@stern.nyu.edu

On Mon, 13 Jan 1997, Francis Heylighen wrote:

> >From: MGeorge@georgegroup.com
> >X-Lotus-FromDomain: GEORGE GROUP
> >To: fheyligh@vnet3.vub.ac.be
> >Date: Mon, 13 Jan 1997 02:09:51 -0500
> >Subject: Ashby's Law of Requisite variety
> >Mime-Version: 1.0
> >
> >
> >
> >I am trying to find books which discuss Ashby's law. Can you suggest some
> >books on Information Theory etc which discuss Ashby's law. I am primarily
> >interested in application to business and economics per the attached file:
>
> ROIC Driven product Rationalization:
> We will probably need to apply this skill in the second quarter of 1997 to
> one or two acquisitions
> 1. What are our past experience levels
> 2.What clients do we now perform this task for
> 3.What holes exist in our methodology
> 4.How are we going to fix any holes
> Please address the issues defined below in your response, and any others
> you may have
> Please give me your thoughts by Jan 15
>
> Dec31,1996: Ashby's law of requisite variety: "The survival of a system
> depends on its ability to generate at least as much variety(entropy) within
> its boundaries as exists in the form of threatening disturbances from its'
> environment"(Krippendorff,p19). Thus we want to increase our product line's
> complexity to match the demands of the market.
> Example: Ford's Model T(1908-1923): Initially there was a huge potential
> demand for cheap utility transportation. Other fancier cars(threatening
> disturbances) had no impact since they did not impinge on the demand.
> However, Sloane's 1922 Chevrolet strategy provided variety which did impact
> Model T demand from above, while used Model T's impacted demand from below.
> Ford's refusal to increase complexity into his product line caused "variety
> to destroy lower variety".
>
> Cost of Variety: Variety exacts a cost in terms of engineering design,
> documentation, production control,inventory, manufacturing engineering,
> capital equipment,manufacturing setups, quality control, purchasing,
> marketing and distribution cost etc. In some large markets, (e.g.
> Bearings), competitors have attained a low cost position by only offering a
> small portion of the spectrum, becoming the low cost producer. thus the
> addition of product lines not only creates the costs specific to the new
> product defined above, but also creates overhead costs which impact the
> previous products. Viewed another way, Chainsaw Al proceeds by cutting 80%
> of the product lines(duplication(3 kinds of paper towels),low GPM high
> invested capital( S.D, Warren) out of a business, outsourcing to the
> max(low ROIC processes), and focusing capital on core competencies in which
> the company can build a competitive cost advantage on a branded product
> which will carry a higher GPM.
>
> ROIC: Viewed another way, Al is "calculating" the ROIC of each product,
> sorting them in descending order, and eliminating those that will not make
> the >5% EVA cut(Scott moved from -3% to +3% in one year). This is merely a
> sudden ROIC product rationalization of a company which had not been managed
> around EVA.
>
> Increasing Variety: So far it sounds like less variety wins. No, its ROIC
> that wins. Ford's persistence with a single product in the face of
> Chevrolet, and his vertical integration at the Rouge plant forced his EVA
> to be negative. After the initial pruning, we want to increase product
> complexity such that the ROIC>5%EVA on the new products (including the cost
> they impose on the other products,ie, they must carry all their own
> costs) . Costs will be lowest if the product line grows like branches of a
> tree, where the new product is a modification of an existing one,
> minimizing the costs above, and using the same marketing channels, and
> going to the same customer base. To the extent this is not true, costs
> will have to be estimated for each item, and will impact the decision to go
> forward
>
>
> ________________________________________________________________________
> Dr. Francis Heylighen, Systems Researcher fheyligh@vnet3.vub.ac.be
> PESP, Free University of Brussels, Pleinlaan 2, B-1050 Brussels, Belgium
> Tel +32-2-6292525; Fax +32-2-6292489; http://pespmc1.vub.ac.be/HEYL.html
>