Scientific Inference Flashcards
Second bundle
Argument
premise»> inference»> conclusion
Includes an initial statement, the premise, intended to determine the degree of truth of another statement, the conclusion
Premise
What we base a conclusion off of
Ex. we have observed 9 red things… conclusion: the next one will be red
Inference
the act or process of reaching a conclusion about something from known facts or evidence
Types of inferences (5)
- DIRECT INFERENCE: “We observe 33% of individuals to be red, therefore, in the population, 33 % of individuals are red.” When you assume that the reality will look the same as the sample group. (inductive)
- PROJECTION: “We observed 9 red individuals, therefore, the next individual we observe will be red.” When you observe a sample group to guess what will happen next. (inductive)
- GENERALIZATION: “We have observed nine individuals of type X. Each of them was red, therefore, we accept the claim that all X are red.” When you generalize something by a hypothesis that everything will be like that. (inductive)
- MODUS PONENS: “Premise 1: if P then Q. Premise 2: P. Conclusion: Q.” We accept a number of assumptions as true, therefore, we accept the implications of these assumptions. (deductive)
- MODUS TOLLENS: “Premise 1: if P then Q. Premise 2: not Q. Conclusion: not P.” We accept a conditional as true, and observe that its consequent is false, therefore, we conclude that H is false.
(deductive)
Deductive Inference
Deductive inferences rearrange current knowledge in such a way that they merely explicate what we already know.
Conclusions from good (“valid”) deductive
inferences and true premises are NECESSARILY true
Inductive Inference
They go beyond information we already have, thus they amplify our knowledge when they are used.
Conclusions from good inductive inferences and
true premises are fallible – THEY MIGHT BE FALSE
! Problems with Inductive inferences
That they might be false, don’t have enough information to be conclusive
Hume’s problem of induction
(An argument against the justifiability of induction. Its premises are:)
- That a given inductive inference rule cannot be justified deductively.
- That the justification of a given inference rule itself consists of an inference from true premises.
- That every inference is either inductive or deductive.
(When inferring I inductively, we must appeal to another (inductive) inference rule
J to justify this induction. But that raises the issue of how to justify J, which would
require appealing to another inference rule K, ….. [infinite regress]»_space;» Consequently, no inductive inference rule can be justified)
! Foundationalism vs Coherentism
F: Identifying the basic claims from which the claims to be justified can be inferred.
C: The claims to be justified from a coherent system with the set of other claims already accepted.
(Both offered by justification)
Hypothesis criteria
- A statement that can be either true or false
- A statement that is not necessarily true or false
- A statement that either has some generality (e.g. “all X in domain D…”), or that is about some unobservable (exclude statements like “this table is red”)
! How to test a hypothesis
- Formulate a hypothesis H
- Deduce observable consequences {Ci} from H.
- Test whether {Ci} is true or not.
- If {Ci} is false, infer that H & {AHj} is false.
- If {Ci} is true, increase confidence in H
Falsification
An event – The observation that an implication of a hypothesis is not true, which by modus tollens then implies the falsity of the hypothesis.
Hypothetico-Deductive (HD) method:
aka, how do you falsify a hypothesis?
- Begin by proposing (unproven) hypothesis.
- They derive observable implications from these hypotheses.
- They test these implications and consequently revise their confidence in these hypotheses.
- If the deduced observable consequences are false, infer that the hypothesis is false.
(uses both induction and deduction)
Karl Poppers falsification
- Conjecture falsifiable hypotheses.
- Seek to falsify these hypotheses with observable evidence.
- Reject any falsified hypothesis as false.
- Never accept any hypothesis as true – only maintain non-falsified hypotheses as so far not rejected.
Falsifiability
Quality of a hypothesis – A good hypothesis has more observable consequences that sets it apart from rival hypothesis.