Parin

Can Inductive Beliefs Ever be Justified?

Introduction
Most of our knowledge that we have about the world we live in extends beyond the scope of our
limited sensory experiences and memory. The way we come up with inferences about such matters of
fact is due to our usage of inductive beliefs.
Overview
In this essay, I will first define and give an example of an inductive inference. Secondly, I will try to
provide some criteria to see if it is justified to believe in inductive inferences. Thirdly, I will analyse
the arguments of the sceptics of inductivism. Fourth, I will provide the counterarguments of the
optimists of inductivism who are generally categorised into three groups: deductivists, pragmatists
and externalists. Fifth, I will try to see what the sceptic’s response to externalism is, if possible.
Finally, I will provide a final summary and detail my suggestions for a resolution to the question of
whether we can be justified in believing in inductive beliefs.
Inductive Inference
An inductive inference:
i) Takes as input previous observations of events and their order
ii) Generates a belief about their relationship
iii) Outputs a belief about unobserved objects or events
An example of an inductive inference would be:
I see the Tom Tower along St Aldates the 1st, 2nd, and 3rd time I walk out of Pembroke.
I believe that if I walk out of Pembroke the nth time and along St Aldates, I will see the Tom
Tower.
Criteria for justified belief
Any sceptical argument that is a general dismissal of empirical knowledge cannot qualify as valid
argument against inductive inference. (Containment Desideratum)
Any argument has to form some basis of an epistemic justification of believing in inductive
Inductive scepticism
The problem with inductive inferences was first raised by Hume 1 who argued that unlike apriori
knowledge (e.g. mathematical concepts, geometry), posteriori knowledge or matters of fact are not
ascertained in the same way; they cannot be distinctly conceived of by the mind and we rely upon
both previous observations and our inductive inference of the ‘law of cause and effect’ to conceive of
them. Then, Hume asks what are the grounds for us to believe that the ‘law of cause and effect’ will
continue to function beyond our sensory experience and memory (i.e. in the future)? We would give
the intuitive answer that experience shows us that cause and effect has worked well so far, as Russell
puts it, that past futures have acted in accordance with the law of cause and effect based on their
respective circumstances, so our current future will also pan out in accordance with the law of cause
and effect. At this point, an objection can be raised. First, that we can see that when one level of

1
David Hume , An Enquiry Concerning Human Understanding, 1748. E. Steinberg (ed.), Cambridge: Hackett,
1993.
inductive inference (e.g. cause and effect) is brought into question, the answer (e.g. experience shows
that cause and effect has worked and will work) also has made an inductive inference as well.
Hume suggests that evidently, the gap between our observing corresponding events and then making a
causal conclusion as to their relationship cannot be explained in terms of reason. Thus he argued, that
this could be explained as ‘custom’ or ‘habit’ of humans, where the repetitive occurrence of a
particular act increases our propensity to expect or behave accordingly. I may suggest that in fact it is
due to our perception of the world as being dominated by cause and effect that itself makes us seem to
ask for a cause of the cause and effect system which is outside of cause and effect, making it seem
incoherent and perhaps even illogical.
In line with this, we can see how, according to Hume, it may be that inductive inference cannot be
reasoned and so believing in them cannot be epistemically justified.
Inductive optimism
Having seen the main inductive scepticist’s argument, let us now consider the variety of responses
coming from the optimists’ camp, who argue that in fact inductive inferences can be justified.
i) Deductivism

The deductivists approach is one that claims that inductive inferences are deductive
inferences (inferences which are entailed by valid premises). They try to justify this
by saying that inferences actually have covert inputs. An example of this would be:

a) Anthony rows for Pembroke

b) Thus, Anthony wakes up early every morning.
The covert input here is that the Pembroke team rows in the early morning everyday .
With the addition of this covert input, we can see how the premises, if valid, clearly
entail the conclusion. In a similar way, deductivists claim that the Uniformity Claim
is a covert input that turn our observational inputs into valid conclusions/beliefs
about the future. An example of this would be that:
a) The first cup of black coffee I drank was bitter. As was the second and third.
b) Uniformity Claim: If the first, second, ….nth cup of black coffee that I drank
was bitter. So the next cup of black coffee that I drink will be bitter.
c) Thus, the next cup of black coffee that I drink will be bitter.
So, showing that the inductive inferences can be interpreted as deductive inferences,
they are thus logically valid (assuming the uniformity claim is valid) and so
deductivists claim this is how to epistemically justify inductive inferences.
One objection that we could pose is that we could challenge the central assumption of
deductivists, that the Uniformity Claim is a valid premiss because we could again
refer back to Hume. He argued that we have no logical reason to believe that the
physical world is uniform as it is innately inductive again to assume that every action/
object are bound by general laws that are constant and uniform.
Another objection to the deductivists approach is that purely by showing an inductive
inference is a deductive inference, it doesn’t really show that it is logically valid. It is
quite likely that we can be sceptical of deductive inferences as well by asking an
infinite regress of why-questions such that it is practically impossible to justify the
validity of even the deductive approach.
 Thus we can see how we can discount the deductivists approach as a method of
justifying the belief in inductive inferences.
Pragmatism
i) An alternative inductively optimistic approach we could use can be pragmatic response to
inductivism. This was proposed by Reichenbach who argued that inductive inferences
were pragmatically viable and is more likely to true relative to all other options of
knowledge. We can say that a method is the best if:
a) x can be the best method of belief in some possible world
b) it is impossible for any other method of belief to be better than x in a possible
world.
This is demonstrated as follows:

ii) either the world is uniform or not

iii) inductive inference is the best possible method of belief if it is so in a possible world and
a) and b) above apply here with respect to inductive inference
iv) If uniformity does hold, inductive inference is the best available method for forming
beliefs about the future
v) If it doesn’t hold, then all methods of belief are ambiguous and so no other method of
forming belief is available.
vi) So inductive inference is the best future belief forming method.
vii) If inductive inference is the best future belief forming method, then it should be justified
as a belief forming method in general.
(Note that this line of argument tries to pragmatically form an epistemic justification of
knowledge.)
An objection to the pragmatic response is that despite it being the best available method of belief
forming, it could be still be a logically invalid/bad method.
For example:
1. Either I have developed clairvoyance powers (no extra evidence) or I have no clairvoyance
powers and I am simply guessing
2. If I have developed clairvoyance powers, then my beliefs are the best method of predicting
the future
3. If I cannot predict the future, then I am just guessing and there is no superior method of
predicting the future, because all other methods are worse than clairvoyance if it works
4. Belief in my predictions is the best method of predicting the future.
Intuitively, believing my predictions about clairvoyance given no extra evidence is unjustified so
dominance arguments that Reichenbach’s uses clearly don’t work. Thus, here too we can see that the
pragmatic response doesn’t yield a justified method of believing in inductive inference.

Externalism
The final inductive-optimistic response that we will analyse is the externalist’s approach, who
challenges the method of justification. Externalism argues that one does not need internal epistemic
justification, purely the fact that the belief forming method is reliable in producing correct
outcomes, then that is sufficient for its justification as a belief forming method. In line with that,
using my first example again here, although the internalist would say that (assuming I am naïve of
inductive inferences being possibly reliable or not and I am naïve of the world being possibly
uniform) I am unjustified in believing that I will see Tom Tower the next day when I walk along St
Aldates. However, as I am I fact in a possible uniform world and inductive inference is possibly valid,
thus the externalist will argue that I am justified in believing so because I will in fact see the Tom
tower as a result of the aforementioned conditions. Thus, it is possible for inductive beliefs, so it is
justified for us to inductive optimists. However, I will object by saying that the same statement may
also give equal credit to skepticists to justify their own position.
Inductive Sceptic’s Response
Considering that the externalist approach has opened a possibility of the sceptic being wrong, the only
way the sceptic can try to counter this approach is by arguing that uniform worlds are impossible,
which is not very plausible itself. Either that, or the skeptic can choose to deny externalism as a
whole, which makes him violate the containment desideratum. Both ways, the optimist has been able
to prove a plausible, but very weak claim to justification. (only argued possibility, not certainty)
Possible resolutions
With the above outcomes of the various schools of thought in mind, it is reasonable for us to take one
of three positions:
a) We conditionally believe in inductive inference, given that we live in a uniform world
b) Inductive inference in fact outputs justified beliefs (seems circular and epistemically
unjustified)
c) We cannot attribute a cause to the cause and effect law that is outside of its sphere as it is
counterintuitive, hence we cannot logically justify inductive inference.
References:

1. Pragmatism Salmon, W., ‘Hans Reichenbach’s Vindication of Induction’, Erkentniss, 1991: 99-
122
2. Externalism Mellor, H., ‘Inaugural Lecture: The Warrant of Induction, in Matters of
Metaphysics. Cambridge, 1988.
3. D.Hume An Enquiry Concerning Human Understanding; Sections 4-5
4. Bertrand Russell (1912), The Problems of Philosophy ch6. Pg60-69