David E Romm

Rhetoric 1152W

Oral presentation 2/6/01

Wesley Salmon, "Scientific Explanation: How We Got from There to Here"

In the article, "Scientific Explanation: How We Got from There to Here" by Wesley Salmon, the author is, at one point, quote, "literally too astonished for words," unquote. Too bad this condition did not persist.

Salmon is trying for a consensus about the philosophy of science. That is, how science is described. On it's own a tricky proposition, since science itself is a description. So we're going to try to agree about how we talk about description. Fair enough. But if you're going to do that, you should eschew the sesquipedalian nomenclature in favor of simple language.

Salmon starts off with the "old consensus", from barely 30 years previous, called the deductive-nominological model, or D-N. As near as I can figure out, this means talking about something that has already happened. Using words, nomenclature, to attempt to describe what it is you can deduce from the scientific experiment. The problem with this is "that not all scientific explanations are of the D-N variety. Some are probabilistic or statistical." In other words, not all experiments yield a definitive one-to-one cause and effect relationship. Some results happen some of the time. This is described by the inductive-statistical model, or I-S. You suspect a causal relationship when some result happens a measurable amount of time under certain circumstances.

As I mentioned in my other presentation about the Pirsig article, what's required is appropriate technology. The proper rhetorical tools for the proper description. Salmon seems confused that arguments with different premises arrive at different conclusions. Why anyone would be confused about this is confusing. Salmon uses longer and less meaningful words as he digs himself deeper into a rhetorical hole.

There's "the problem of ambiguity of I-S explanation", which is no problem at all. Statistics are a major component of the scientific method. If you don't know how to use statistics, don't become a scientist, and don't pretend to be able to describe science. Better is the "principle of essential epistemic relativity of I-S explanation", which, I think, means that statistical explanations are sometimes true but not always.

Salmon finds no resolution between these differing types of explanations. Therein lies a question: Why does there need to be "a basis for a new consensus"? The methodology of science is wide, and the deterministic nature of physics should be described differently than the probabilistic nature of psychology. For that matter, Salmon doesn't even mention that Quantum Physics proposes that even deterministic science is probabilistic.

I think Salmon is using the language very poorly. He's claims to be trying to achieve consensus in the philosophy of science, when he's merely playing with the rhetoric of science. He's using five-dollar words to obfuscate fairly simple problems. Nice that he has a big vocabulary, but it gets in the way.

Having said all that, let's relate some of the points from this article with previous articles in this course. Pirsig would say that it's unnecessary to need to know the philosophy of science at all. Motorcycle repairmen just fix the bike without talking much about it at all. Bronowski is starting at the other end of the scientific method, using the imagination to create experiments. He really doesn't go into post-experimental descriptions at all. Popper is similarly unconcerned with descriptions of the methods, but only counting experiments that test falsifiability. Ziman says that science is "a consensus of rational opinion over the widest possible field". He is more concerned with science as a philosophy, not the philosophy of science. Frankly, there aren't any questions in this essay that Ziman didn't answer better.

Back to main Rhetoric Page


romm.org Portal PageFamily Pictures & e-mail | Avrom Romm obituaries and text files
Picture Galleries | Shockwave Sound Files | Shockwave Home Page