Carl Hempel: “Studies in the Logic of Explanation”

 

Internet Link:

Internet Encyclopedia of Philosophy: Theories of Explanation

 

Background:

Hempel is a Logical Positivist, keen to give an empiricism-friendly analysis of the otherwise suspect notion of explanation.  It is apparently suspect because it looks metaphysical, that is, it looks like a good explanation must make reference to nature as it really is, which is something we can never know for certain.  Alternatively, it could be divorced entirely from reality and just couched in terms of what the particular audience finds convincing.  Thus, anthropomorphic explanations of the world are more appealing, and appeal to Homer’s Gods was the best kind of explanation for the time, because it enhanced “understanding”.  Hempel wants to avoid both these extremes.  He avoids metaphysics because his conception of explanation is epistemological: that is, the standard of a good explanation is not determined by some unreachable reality, but by good epistemological rules.  And he avoids relativism by coming up with a rigid formal structure that all explanations must exhibit, and in reference to which we can distinguish real explanations from pseudoexplanations.

 

I. Elementary Survey of Scientific Explanation [91]

 

1. Some Illustrations [91]

Event: mercury thermometer is placed in hot water, and before the mercury rises, it briefly drops.  Why?

Answer in terms of the comparative expansion rates of mercury and glass, and the fact that the heat reaches the glass first:

Thus, the event under discussion is explained by subsuming it under general laws, i.e., by showing that it occurred in accordance with those laws, in virtue of the realization of certain specified antecedent conditions. [92]

Example 2: oar in water

Thus, here again, the question “why does the phenomenon occur?” is construed as meaning “according to what general laws, and by virtue of what antecedent conditions does the phenomenon occur?”

Thus Hempel has examined two examples of good explanation and found a general underlying structure, which he proceeds to outline:

 

2. The Basic Pattern of Scientific Explanation [92]

Two components of an explanation:

  1. the explanandum (sentence describing the thing to be explained [but not the thing itself, perish the thought, that would be metaphysics!])
  2. the explanans, consisting of
    1. antecedent conditions (like the state of the glass and mercury before submersion into the boiling water)
    2. general laws

An explanation is a logically valid argument with the explanans as premises and the explanandum as conclusion.  Thus, what an explanation does is show how the thing-to-be-explained follows logically from initial conditions plus general laws.  That is, the event is something that would have been expected given our knowledge of the way of the world before the event plus the laws.  This is why Hempel’s explanations are called Deductive (logical deduction)-Nomological (to do with laws) explanations.  It also shows why he thinks that explanations and predictions have exactly the same structure – every explanation could have functioned as a prediction before the fact.

 

Logical conditions of adequacy:

(R1)                      The explanandum must be a logical consequence of the explanans

(R2)                      The explanans must contain general laws, and these must actually be required for the derivation of the explanandum

(R3)                      The explanans must have empirical content (i.e., must be testable)

Empirical condition of adequacy:

(R4)                      The sentences constituting the explanans must be true

R4 sounds suspiciously metaphysical, but Hempel needs it to avoid relativism.  “Highly confirmed” would be the alternative to “true” but that would mean that laws that appear highly confirmed by their day’s standards would have been good explanations at the time and now they’re not.  But “this does not appear to accord with sound common usage” whereby good explanations are always sound.

 

Incomplete explanatory arguments are like enthymemes – that is, arguments with missing premises.  These are okay (that is, the FULL explanation model is just an ideal, not every good explanation has actually to be in that form, it just has to be presentable in that form by supplying uncontroversial suppressed premises) so long as the missing parts are legitimate laws (e.g., sugar dissolving in tea [95]) but not otherwise (works of an artist “explained” in terms of neuroses)

 

3. Explanation in the Nonphysical Sciences.  Motivational and Teleological Approaches.

Examples [96]:

Economic/Sociological: 1946 severe fall in the price of Cotton

Linguistics: one word for “bee” in Southern France, but lots in the North

 

Hempel argues that all good explanations in the “Nonphysical” sciences can be put into this format.  This is naturalism, that is, the idea that the non-physical sciences should follow the pattern of physics.

Objections to naturalism:

  1. Human behaviour is unique and irrepeatable (sic)
    RESPONSE: this is true of physical events too: each event is unique in its possession of all its features (particularly the time it occurs) but several of its features may conform to general laws, and in this sense it can be explained.  Thus, human events can be explained too, even if they’re unique.
  2. Behaviour of individual humans depends on personal history
    RESPONSE: this is no barrier to generalizations, as they can take account of the history of individuals.  E.g.: “For people of x history in y situations…”
  3. Human behaviour is teleological (goal-oriented) and thus not subject to the usual causal generalizations.  The explanation of human behaviour has to take their goals into account rather than simply their muscle twitches.  But motives and goals are unobservable.
    RESPONSE: (a) There are general laws of human behaviour: they are in terms of beliefs and desires: “If x believes y and desires z, she will perform action A”
    (b) there are unobservables in physics too, but those unobservables (e.g. electrons) have observable consequences, and so do motives

 

Important distinction: understanding in the psychological sense vs. the theoretical (or cognitive) sense.  Teleological explanations help with the former because they are anthropomorphic and therefore provide familiar explanans.  BUT: familiarity is neither necessary (a body falling is scientifically explained by the law of gravity, which is less familiar than the event it explains) nor sufficient (example of “entelechy or vital force” [100-101])