Beruflich Dokumente
Kultur Dokumente
This schema rests on four premises; the explanans only provides a scientific
explanation of the explanandum iff:
(i) E is a logical consequence of the conjunction of the explanans
(ii) E does not follow from any proper subset2 of the explanans
sentences
(iii) The explanans sentences must have empirical content3
(iv) The explanans sentences must all be true. (Taken from Ruben, 1990)
-1-
Nick Fletcher PF303
For example, a scientific explanation of why the Wasps beat Bristol Rugby
recently could have as initial conditions that that the Wasps have a stronger and
heavier pack (I1) and faster and more skilled backs (I2). Statistically, as it is more
likely that the team with a stronger, heavier pack and faster, more skilled backs
tend to win in rugby, this probabilistic law can be used to provide a scientific
explanation of the win. The probabilistic nature of this law is highlighted by the
fact that the score was 32-30, with the winning try being made in the 80th
minute; the law is clearly probabilistic rather than exceptionless as things may
well have actually gone the other way.
These models suggest that there is a strong symmetry between
explanation and prediction Indeed, if the model holds then the only difference
between the two is that prediction occurs before the event has occurred and
explanation afterwards (Papineau, 2005). In this way, the model appears to
-2-
Nick Fletcher PF303
-3-
Nick Fletcher PF303
shadow, and therefore can also be used to scientifically explain it. This is because
there is a sort of causality to such coexistences. If the shadow did not exist in the
example, perhaps due to someone shining a torch where the shadow should be, it
would not affect the existence of the pole; however, if the pole were to cease
existing, the shadow would cease existing also. The existence of the shadow
depends entirely on the existence of the pole, in a way it is caused by it, whilst
the existence of the pole depends in no way upon the existence of the shadow.
Whilst this idea does not fit with Hempel's beliefs about causation, that causation
must involve a succession in time, using this different view of causation does not
cause problems in the DN model. Thus, (b) holds and the symmetry thesis can be
used as a gage as to whether an explanation is truly scientific.
It could be argued that the addition of causality to scientific explanation
causes problems in terms of constitutive explanations and teleological
explanations. A constitutive explanation could take the form of:
L: Water is H20,
I: The liquid in the glass is H20
E: Therefore the liquid is water5
This is indeed a truly scientific explanation of why the liquid in the glass is water,
but it is not clear where the notion of causality comes into it. However, it could be
argued that the fact that the liquid in the glass is constituted by H20 causes it to
be water, and whilst this requires a slightly broader view of causation6 than is
held by most philosophers, it has intuitive value and serves to allow such
explanations to rightfully be called scientific.
Teleological explanations concern why a certain purpose has been
achieved or end has come about, for example ‘why are polar bears white?’,
‘because it helps them to hunt with more success.’ Whilst this is clearly a
scientific explanation, it is not clear how the polar bears improved hunting skills
can have caused them to be white. However, the explanation can be re-phrased
thus:
L: Evolution means that animals with the best characteristics for survival are
more likely to survive longer an thus produce more offspring,
I: Polar bears live in snow-covered lands,
E: Polar bears have evolved white fur to help them hunt
-4-
Nick Fletcher PF303
All that has occurred here is a law has been introduced that, in conjunction with I,
truly explains E scientifically by creating a causal link between I and E. E is shown
to be, in a way, caused by both L and I, and causation remains valid as a
component of scientific explanation.
Thus, from his covering-law model of explanation, Hempel has been able
to produce two models of scientific explanation, DN and IS. One of these models
is deductive and the other inductive as, just like any argument, scientific
explanation can occur either deductively or inductively. Between them these
models seem able to cover all instances of scientific explanation, albeit after
taking a broad and quite intuitive view of causation; including those that involve
coexistent laws and successive laws, as well as straightforward, constitutive or
teleological explanations. Furthermore, mere explanations can be distinguished
from truly scientific explanations via the symmetry thesis. To conclude, a
scientific explanation is7 one that takes the form of either a deductive or
inductive argument involving laws or initial conditions that both entail and are
causally related to the explanandum.
1
There is also a more relaxed version of the covering-law model that does not
insist on this deductive element, the statistical-relevance model, which shall be
discussed later in this essay.
2
i.e. there are no surplus explanans that are not necessary for the explanation.
3
As noted by Ruben (1990), this point is fairly redundant as the explanandum is
necessarily an empirical fact, meaning that (i) ensures (iii) is necessarily fulfilled.
4
This is not an exact definition, but serves the purposes of this essay.
5
The terms ‘H20’ and ‘water’ can be interchanged according to the information
-5-
Nick Fletcher PF303
Bibliography:
-6-