< < <
Date Index
> > >
Fw: Affective measures in the social sciences produce more ideologic agitprop...
by Mike Alexander
16 September 2002 21:42 UTC
< < <
Thread Index
> > >
John Landon writes:
I was referring more generally to Very Large Scale historical 'evolution', and there we have to take everything into account from high to low. And the complexity is too vast for any simple model.
Very complex phenomenon can sometimes be described adequately using very simple models.  For example, consider the growth of micro-organisms.  This is an enormous complex process involving hundreds of individual physiological processes, which interact with dozens of environment variables.  Yet under certain conditions the growth rate can be described reasonably well by the Monod equation:
dX / dt = X * vS/(S+K)  where X is the microbial mass, S is the concentration of the substrate (food) and v & K are constants.
The certain condition is when growth is limited by the availability of food as opposed to limitations of air or water, or the accumulation of toxins.  Fernand Braudel notes that at certain times in history, e.g. early 16th century, (also early 14th) real wages fell to a level where the amount of grain they would purchase barely matched the caloric needs of a person.  In other words, lower wages would be incompatible with life.  And indeed wages did not fall, rather, population did, and wages rose.  So human populations can become substrate-limited just as microbial populations.
But there are more interesting responses of microbial populations to environmental consequences of their own behavior.  In my Ph.D. research on alcohol production from xylose (a wood-derived sugar), I noted cyclical behavior of the yeasts under certain conditions.  I would do experiments in which I initially produced a stably-growing population of yeasts in a continuous fermentor.  A continuous fermentor has a feed and a discharge which maintains a constant volume.  At steady-state the culture is limited by the rate at which substrate is fed into the fermentor--thus growth rate can be manipulated as an independent variable.  This initial population would be exclusively performing respirative metabolism (they consumed sugar, growing new biomass, and releasing carbon dioxide and water as waste products like we do).  I would then suddenly shift them to conditions of partial oxygen deprivation in which they no longer had sufficient oxygen to metabolize all the sugar available.  This would kill higher animals like humans, of course.  But yeast are facultative, that is, they can do both respiration and fermentation.  So after a lag of a few hours the yeasts would metabolize via fermentation (producing alcohol) that sugar they could not metabolize via respiration (using oxygen).  The result would be accumulation of alcohol until a new steady-state was reached, in which biomass was produced using a combination of fermentative and respirative that was controlled by the availability of oxygen. 
In some experiments, those in which I started with a large population of yeasts and tried to make high concentrations of alcohol, I found that a steady state alcohol concentration was not achieved.  Instead the concentration of ethanol (and every thing else) oscillated between high levels and low levels.  The high levels were higher than (and the low levels were lower than) the steady-state levels I had obtained in other experiments with smaller populations.  What was happening was accumulation of alcohol (which is toxic) was adversely impacting the growth and fermentation capabilities of the yeasts. This is a commonly observed phenomenon, but even if it happened, one would still expect steady state conditions to be reached.  The nature of these steady states would depend on both the alcohol concentration and the oxygen supply.
But instead I observed cyclical behavior.  After playing around with the data a bit I noticed that one could model the effect of alcohol on growth and fermentation if you used the alcohol concentration of the past rather than the present in the model.  There was a "memory effect".  Alcohol exerted a deleterious effect on microbial life, which took time to manifest.  During this time the concentration would continue to accumulate.  At some point the past alcohol levels finally exerted an effect strong enough to stop increasing present alcohol levels, but by this time the level was much higher than the problem level.  So now alcohol production began to fall in response to the still-rising concentrations of the past.  By the time the alcohol production level stopped falling, the level had reached a present level way below the problem level, and so up the concentration would go. The lag time required to model this behavior was the "generation time" of the population.
I hypothesize that late Medieval and Early Modern human populations did something similar.  But instead of waste products, the driver here was lagged response of fertility to food supply.  I start with a time when food supply is growing at a faster rate than population.  People coming of age at such times will, on average, find it easier to establish a household and to have a family.  They should experience greater lifetime fertility.  Because of this increased fertility the population a generation down the road is going to contain more people of breeding age than the current population.  Unless the food supply grows at a faster rate, this larger population is going to fast a less satisfactory relation between food supply and mouths to feed.  They will find it harder to establish a household and start of family.  They will experiment reduced fertility. 
Thus we might expect to see oscillating  fertility and relative food supply.  When fertility is low relative food supply is high and their should be fewer famines and food prices should on average trend downward. When fertility is high, relative food supply is low and their should be more famines and food prices should trend upwards.  The oscillations should show a time constant of (human) generational length, which peaks and toughs spaced two generations apart (50-60 years).  This cycle in prices is called the Kondratiev cycle.  Evidence for this model would be a cyclical behavior of  famine frequency, with frequency highest at the point of the fertility cycle at which food supply relative to population is the least favorable and prices should average the highest (i.e. the Kondratiev peak).  A statistically significant pattern does appear in these variables as I describe in my Kondratiev book.  Other evidence is also presented.
This is an example of how a natural science approach (that is not physics-based) can be applied to human behavior.  Some of the long term trends you mention follow naturally from this sort of a mechanism--given certain constraints imposed by geography.  I can get into these later, if people are interested.
Mike Alexander,  author of
Stock Cycles: Why stocks won't beat money markets over the next 20 years and
The Kondratiev Cycle: A generational interpretation
----- Original Message -----
Sent: Sunday, September 15, 2002 10:12 PM
Subject: Re: Affective measures in the social sciences produce more ideologic agitprop...

In a message dated 9/15/2002 9:06:53 PM Eastern Daylight Time, malexan@net-link.net writes:

Predictions do not have to be precise to be useful. A statistical forecast of the sort that there is a Y% probability of A happening, but only a Z% probability of B happening can be useful, even if the exact state (A or B) is not predicted.

I would certainly accept that. I note you are dealing with stock markets. That is one phenomenon. I was referring more generally to Very Large Scale historical 'evolution', and there we have to take everything into account from high to low. And the complexity is too vast for any simple model. We find even the 'evolution' of philosophy of science, and suddenly it gets hard to compute the evolution of the very tools we are using.
The form, by the way, of Marx's thinking both conforms to this, but at the same time is often taken in the wrong sense. One can take a sort of Popperist viewpoint of such things, OK. But the sense of Marx is different from 'causal' prediction.
Looking at world history we could predict over the long term the economic domination of events, yet again over the long term a mysterious long term trend toward 'freedom'. This, I think, is the point of Marx's predictive thinking. But the status of prediction is different in both cases.
Part of the problem (note my Rousseau post here) is that men have to shake themselves lose and assume their 'freedom'. Otherwise the myth of historical continuity will cast its spell, and what has great momentum will seem inevitable, how things should be because they must be.
Take the example of slavery. The long term trend up to the medieval period was the slow but steady progression of the form, very predictable worsening over the long term. Yet trends toward freedom are emerging, first as small scale influences, then in the modern a conclusive development. How and why does it happen?
I think this is altogether different from standard predictive models. But the example of the rise of the phenomenon of bourgeois 'freedom' was what on the one hand whetted the appetite and on the other lead to a prediction that 'real freedom' in some socialist version was 'inevitable'. It is an elusive question, and I think we can miscontrue Marx here and run afoul of the Poppers who rightly note that a deterministic law of history here miscontrues the question.
But the fact remains that the explosive rise of the modern saw the convergent collision of the two streams, the long term domination of slavery amplifying, and yet the suddent trend toward freedom that resulted in what we now take for granted, the abolition of slavery. These issues seem obvious, but to correct model these two different types of processes makes for a type of model different from the purely causal. In that sense Marx's historical materialism has too often been miscontrued.
This type of distinction requires separating the analysis into the characteristic levels of micro and macro, but applied to some 'idea for a universal history'. And we can only produce that looking backward in the past, for the nature of the future has to be created freely. It is a paradox that doesn't easily lead to correct analysis.

John Landon
Website on the eonic effect
< < <
Date Index
> > >
World Systems Network List Archives
at CSF
Subscribe to World Systems Network < < <
Thread Index
> > >