Why do we keep talking about levels of explanation? Where do these levels come from anyway? What's a level?
O'Neill et al. say it's all about rates -- all about temporal scale.
-
Nature is processes.
-
Processes occur at different rates.
-
Processes are linked to each other. As modelers, we often think of processes as being (or represented by) variables.
-
One process is linked to another. One variable responds to the value of another. Same thing.
-
A variable responds to change in another only if the variation in one is on a similar temporal scale to the scale of variation in the other.
-
Levels: higher level is slower rate -- is larger temporal scale. Lower level is faster rate -- is smaller temporal scale.
-
Process "perceives" a slower process as an average of values. Process perceives a faster process as a constant.
-
Think of sci-fi in which a character is pulled into a different temporal scale. A faster one. S/he is suddenly invisible to everyone else, and everyone else becomes statuelike.
-
Examples: prices are set on large temporal scale, while consumers' decisions are made on small temporal scale. Prices clearly come about because of consumer choice, but prices don't respond to little blips in consumer activity. Prices respond to some kind of integrated aggregate of all the values over a longish time. At the same time, consumer choice isn't dependent on all the changes in price -- just in whatever the price is at the present time. As far as consumer choice is concerned, price might as well be a constant.
But we should bring all this back to models and explanation.
Hempel and Oppenheim mention emergence. Some macroscopic properties of water are easily derived from the propoerties of water's microscopic components. Example: the mass of a liter of water. Some macroscopic properties cannot be derived from the properties of the components. Example: transparency. Does this mean that we can't explain transparency?
But doesn't that question depend on us thinking that lower-level processes cause the higher-level phenomena?
What does cause mean anyway? Well, we won't get into that. We'll just use a layperson's intuitive operational definition, or test: A causes B if getting rid of A prevents B. Bad philosophical definition, but good everyday definition -- and it's the definition that underlies almost all of experimental science. Example: not putting Mystery Coronavirus into monkey's nostrils prevents the monkeys from contracting SARS. For the time being, we say that the virus causes SARS.
Last time, Hannah said that if you tried to explain the Yankees' success in World Series by looking at quarks, by the time you get macroscopic again there might not even be baseball. This sort of statement requires there to be causation between levels, just like we think of causation as occuring in time. Compare to Stephen J Gould's question, if you were to rewind and then play back the tape of the history of life, would we still get the same organisms today? Gould says no, like Hannah suggests that moving down to quarks and then up to baseball teams, you might not get baseball teams at all. Levels have causal connections like the ones we usually associate with events.
How do the levels fit in with the three things Levins's generality, precision, and realism?
| high level | low level | |
| realism | lower | higher |
| precision | higher | lower |
| generality | higher? | lower? |
