As researchers working in government, influencing service design, we need to know that our research is methodologically sound, our research findings are grounded in empirical data and our recommendations are logically derived.
'Understanding arguments, reasoning and hypotheses' is the first in a series of 5 short courses, covering introduction courses to various aspects of methodology in research, from the use of grounded theory in discovery research, to hypothesis testing and sampling in more experimental research.
In this course, you'll learn:
About arguments
- what we mean by an argument
- how to identify a valid/invalid argument
- what we mean by premises
- what validity and soundness of arguments mean
About reasoning
- what is deductive reasoning and where do we use it
- what is inductive reasoning and where do we use it
- what is abductive reasoning and where do we use it
About hypotheses
- what is a hypotheses and a null hypothesis
- how do we test them
10. Logic is the study of arguments,particularly
of valid arguments.
A logician is a person who worries about
arguments and they can be from anywhere
science,technology,politics…etc.
How can we tell good arguments from bad
arguments,from assessing the sentences
themselves.
Fathers of classical
logic:
Frege (right)
Leibniz (bottom right)
22. (SOMETIMES REFERRED TO
AS BELIEFS)
Propositions can be thought of as simple declarative sentences
ü It is raining
ü Elephants like peanuts
ü Paris is the capital of Germany
A proposition is something which can be either true or false.
In logical jargon, we’d say that a proposition must have a truth-
value.
25. Sentences like the below aren’t propositions.
‘What’s the time?’
‘pass me the salt’
ØWhy?
26. Sentences like the below aren’t propositions.
‘What’s the time?’
‘pass me the salt’
ØWhy?
A useful tool for ‘testing out’a proposition is to stick the
phrase‘It’s true that…’at the start of the sentence,and see
whether it makes sense.
‘It’s true that “it’s raining outside’’makes perfect sense!
29. A counter-example is a statement which disproves the assertion
made by a given proposition.Take the following proposition:
‘All birds can fly’
Can you think of a counter-example?
30. The beauty of arguing by counter-examples is that,if you can
find just one, then you can be sure that the form of the argument
is invalid.
Note:The argument itself might be intuitively ‘good’,but if you
discover a counter-example,it can no longer be classed as
‘logical’.
33. § Consistency applies to sets of propositions.
§ For a set of propositions to be consistent,then,there must be at
least one situation in which they could all be true together.
§ If that hypothetical situation is impossible—that is, if the
propositions in our set couldn’t all be true at the same time—
then we say the set is inconsistent.
34. § A set of propositions is consistent if, and only if, it’s possible that
all those beliefs could be true at the same time.
§ A set of propositions is inconsistent if, and only if, it’s impossible
that the beliefs could all be true at the same time.
42. Philosophers (mathematicians,scientists,and so on) use the
term‘argument’in a precise and narrow sense.
An argument is made up of‘propositions’ which either act as the
premise(s) or the conclusion.
Here’s an example of an argument:
All men are mortal
Socrates is a man
Therefore, Socrates is Mortal
43. Philosophers (mathematicians,scientists,and so on) use the
term‘argument’in a precise and narrow sense.
An argument is made up of‘propositions’ which either act as the
premise(s) or the conclusion.
Here’s an example of an argument:
All men are mortal Premise (universal)
Socrates is a man
Therefore, Socrates is Mortal
44. Philosophers (mathematicians,scientists,and so on) use the
term‘argument’in a precise and narrow sense.
An argument is made up of‘propositions’ which either act as the
premise(s) or the conclusion.
Here’s an example of an argument:
All men are mortal Premise (universal)
Socrates is a man Premise
Therefore, Socrates is Mortal
45. Philosophers (mathematicians,scientists,and so on) use the
term‘argument’in a precise and narrow sense.
An argument is made up of‘propositions’ which either act as the
premise(s) or the conclusion.
Here’s an example of an argument:
All men are mortal Premise (universal)
Socrates is a man Premise
Therefore, Socrates is Mortal Conclusion
46. Arguments attempt to expand our knowledge.If you have
good reason to believe an argument’s premises,then a well-
structured argument will give you good reason to believe the
conclusion too.
In logic,the argument is the smallest individual piece of
reasoning.(If arguments are the ‘atoms of reasoning’,then
propositions are the sub-atomic particles.)
51. Validity refers to arguments:arguments are either valid or
invalid
A valid argument is one where it is impossible that the premises
all be true and the conclusion false.
v All humans breathe air
v I am a human
v Therefore,I breathe air
52. Another way to think of this is as follows: if the argument’s
premises are all true, then the conclusion must be true also.
If there is even one situation in which the argument’s premises
are all true, but its conclusion is false, then we say the argument
is invalid.
59. A valid argument whose premises are all actually true is called a
sound argument.
§ Note:Every sound argument is a valid argument;it is not
possible for an argument to be invalid and sound!
62. § Arguments can only be valid or invalid. Propositions can only be
true or false.
§ There is no such thing as a ‘true argument’ or ‘false argument.’
Likewise, there is no such thing as a ‘valid belief’ or an ‘invalid
belief.’
§ A valid argument whose premises are all true is called a sound
argument.
63.
64. So far,we’ve said that an argument is valid if
(and only if) the conclusion follows necessarily
from the premises.
However, we can hone this idea a little more by
introducing the notions of deduction and
induction.
69. We have already covered deduction,when we covered
arguments;a deductive argument is where the premises supply
all the information we need to see that the conclusionis true.
v If it’s a raven,then it will be black
v It is a raven
v So,it will be black
Here,the premises of the argument supply all the information
we need to say whether the conclusion is true or false.
76. An inductive argument is one which moves from
observations to a universal statement.
It takes the following form:
Raven no.1 is black
Raven no.2 is black
Raven no.3 is black
…
Raven no.n is black
So, all ravens are black!
Francis Bacon
1561-1626
Philosopher, scientist
79. In deduction,the truth of the conclusion
follows from the truth of the premise.But, in
induction,the truth of the conclusion is not
guaranteed by the truth of the premises.
The philosopher Ian Hacking calls them
‘risky arguments’for just this reason.As a
result,inductivearguments can be very
good arguments—but they can never be
valid!
80.
81. WHICH OF THESE ARGUMENTS ARE
INDUCTIVE, AND WHICH ARE DEDUCTIVE?
85. Fans of Sherlock Holmes might recall that
character’s frequent references to the
‘science of deduction’.
ØIs this use of the word‘deduction’
correct?
Pixabay.com. Creative Commons license CC0.
86. Actually, what Sherlock is usually doing could be described as
abductive reasoning.
Abductive reasoning, then, can be thought of as inference to
the best explanation.
A lot of what you’ll do as researchers will involve some form of
abductive reasoning.
87. Suppose I come home and find that all the milk that I had in the
fridge has disappeared.How did this happen?
Any number of hypothetical situations is possible: perhaps a
thirsty burglar broke into my house! Perhaps,for some
inexplicable reason,the fridge became very hot and all the milk
was evaporated.However,much more likely than these is that I’d
drank all the milk, and had simply forgotten.
88. We should note,however,that,like inductive arguments,
arguments based on abductive reasoning carry an element of
risk. But, just because an argument isn’t valid in the narrow sense
we’ve described doesn’t mean that it’s a bad argument!
89.
90. §
Image sourced and adapted from
https://bookofbadarguments.com. Ali Almossawi. Creative
Commons BY-NC license
92. Affirming the consequent (i.e. the
conclusion) is a formal fallacy
which takes the form:
v If P --> Q
v Q
v Therefore,P
Image sourced and adapted from
https://bookofbadarguments.com. Ali Almossawi. Creative
Commons BY-NC license
93. The quantifier-shift fallacyis a logical fallacy in which
the different quantifiers used in a statement get mixed
up.
‘Every event has a cause.So,there must be one cause
for every event.’
(We’ll return to this one!)
95. This fallacy is committed
when one forms a
conclusion from a sample
that is either too small or
too unique to be
representative.
Image sourced and adapted from
https://bookofbadarguments.com. Ali Almossawi. Creative
Commons BY-NC license
96. When one event is believed to have caused by
another because of their co-occurrence or where
one event is seen to have preceded another.
97. A false dilemma occurs
when only limited options
are presented,despite the
fact that at least one other
option is possible.
“Your either with us,or with
the fanatics”
Image sourced and adapted from https://bookofbadarguments.com. Ali Almossawi.
Creative Commons BY-NC license
98. The fallacy of equivocationis an
informal,semantic fallacy where
the a term is used which has
more than one meaning (but the
meaning which is intended is
not made clear).
It makes for a lot of our British
jokes…
“The sign said "fine for parking
here",and since it was fine,I
parked there.”
Image sourced and adapted from https://bookofbadarguments.com. Ali Almossawi.
Creative Commons BY-NC license
99. The fallacy of composition
involves attributing a certain
property to a set of things,after
observing that each individual
member of the set has that
property.
And,the fallacy of division,
involves attributing a certain
property to members of a set,
after observing that the set itself
exhibits that property.Image sourced and adapted from https://bookofbadarguments.com. Ali Almossawi.
Creative Commons BY-NC license
100. A slippery slope
argument attempts to
discredit a proposition
by arguing that its
acceptance will
undoubtedly lead to a
sequence of events,one
or more of which are
undesirable
Note the probability of that
eventuality can sometimes be
infinitesimally small!
P(0.2) x Q(0.1) x R(0.6)…
Image sourced and adapted from https://bookofbadarguments.com. Ali Almossawi.
Creative Commons BY-NC license
101. A.K.A. the existence of
[God, aliens, you fill in the
blank] argument
This kind of argument
assumes a proposition to be
true simply because there is
no evidence proving that it
false.
Hence,absence of evidence
is taken to be evidence of
absence.
104. § Propositions are entities that can either be true or false
§ Arguments are sets of propositions containing some premises and a
conclusion
§ Propositions are said to be consistent when they can all be true
together
§ Arguments are valid when it is impossible that their premises be true
and their conclusions false.
106. Propositional logic (the logic of propositions) has five different
‘connectives’ for linking sentences up.
You’ll probably be familiar with all of them already: they
correspond (roughly) to the English words ‘and’, ‘or’, ‘implies’, ‘if
and only if’, and ‘not’.
Let’s take a look at these more closely.
107. P and Q be two arbitrary English sentences.
P might be‘it’s raining’,
while Q might be‘it’s wet outside’.
Now suppose we join these sentences up with the connective ‘and’
to form the new sentence ‘P and Q’.
What’s different about this new sentence?
108. The difference is that the truth of the new sentence depends on
the truth of each of the parts represented by the sentence-letters
P and Q, respectively.
If P is ‘It’s raining’ and Q is ‘It’s wet outside’…
then for the new proposition ‘It’s raining and it’s wet outside’
to be true, both of the ‘smaller’ sentences must be true in turn.
109. We can represent this idea
in something called a
Truth-Table.
Note,then,that the
proposition ‘P and Q’ (‘It’s
raining and it’s wet
outside’) is true only when
P is true and Q is true.
P Q (P and Q)
T T T
T F F
F T F
F F F
112. This one is the easiest of the
logical connectives. For any
sentence,P, adding the word‘not’
simply reverses the truth-value of
P.
To assert that ‘It is not raining’ is
just to say that the proposition ‘It
is raining’ is false.
P not-P
T F
F F
113. Suppose we postulate that P
implies Q.What does this
actually mean?
Going back to our natural
language equivalents,we might
say that ‘It is raining implies that
it is wet’.
Another way to say this is ‘If it is
raining, then it must be wet’.
P Q (P implies Q)
T T T
T F F
F T T
F F T
114. This has unintuitive implications. Suppose we have an implication
like ‘If pigs can fly, then I will be king.’
In this case, both parts of the implication are false.
But, consult the truth-table:if both parts of the implication are
individually false, then the implication as a whole is true.
115. If you find this one hard to stomach,then don’t worry: you’re not
alone!
Just remember that we’re just interested in saying when a certain
state of affairs must entail another,and not in whether or not
either of those states of affairs is actually possible.
116. In essence,it’s just another way of
saying‘P implies Q and Q implies
P.’
Another way to write this is P
implies Q if and only if, Q implies
P.
P Q (P if and only if Q)
T T T
T F F
F T F
F F T
117. Knowing the truth-tables provides you with a useful method for
recognising when propositions,hypotheses,etc., are legitimate and
when they are not.
For instance,knowing that the presence of fire implies the presence
of smoke does not necessarily give you any reason to believe that
the reverse holds too.
119. Another key insight of modern formal logic is into how words like
‘all’, ‘some’, or ‘none’affect the truth of a sentence,or the validity
of an argument.
Words like these are known as quantifiers. There are 2 types of
quantifier.
120. ∀(an upside down ‘A’).You can read this symbol as‘for all’ or
‘everything’, etc.You might like to think about statements
involving this quantifier as being always true or always false.
∃(backwards‘E’), and it can be read as ‘there exists’,‘there is at
least one’,‘some’,‘many’, etc.Basically, it’s anything other than
‘all’. Conversely,you might like to think about statements
involving this quantifier as sometimes true, and sometimes false.
121. People mix them up all the time! (Quantifier shift fallacies)
ØLooks plausible?
‘Every event has a cause. So, there must be one cause for each
event.’
ØWhat about this one?
‘Everyone has a mother.Therefore,someone must be the mother
of everyone’—this is an obviously invalid assertion!
123. v If P --> Q
v P
v Therefore,Q
§ EXAMPLE: I know that if it rains, it will be wet outside. I also
known that it is raining now.
Thus, by modus ponens, I am logically justified in inferring that
it’s wet outside as well.
124. (SISTER OF MODUS PONENS)
v If P --> Q
v -Q
v Therefore,-P
§ EXAMPLE: Suppose I know that if it’s raining,it will be wet
outside.However,upon looking outside,I find that it is not
wet outside.From this, I am justified in inferring that it’s not
raining either.
125. (PROOF BY CONTRADICTION)
this argument form allows us to infer one
proposition,P,by showing that its negation,not-P,
leads to contradiction.
§ EXAMPLE: Galileo’s proof of the law of falling
bodies
(which says that the distance travelled by a falling
body is proportional to the square of the time).
126.
127.
128. Remember propositions?
A hypothesis is a proposition about a state of the world…
E.g. Plants require light for photosynthesis
or more advanced:the level of light plants require for
photosynthesis is proportionate to the rate of photosynthesis
129. We have done lots of observation and we have a hunch about a
causal mechanism
X causesY or X is a cause and affectsY to some unknown
extent Z
We want to test whether our hunch is true!
130. To test the hypothesis, we construct the null hypothesis (which is
the exact negation of the hypothesis).
E.g.
Hypothesis: Plants require light for photosynthesis
Null Hypothesis: Plants do-not require light for photosynthesis
131.
132. WHY DO WE TEST THE NULL HYPOTHESIS
AND NOT THE HYPOTHESIS?
133. Let’s suppose we test the hypothesis.
vIf the theory is correct, it implies that we could observe
Phenomenon X or Data X.
vX is observed.
vHence,the theory is correct.
What’s wrong with this?
134. Let’s look at a similar example..
vIf Jefferson was assassinated,then Jefferson is dead.
vJefferson is dead.
vTherefore Jefferson was assassinated.
It’s invalid and a fallacy. Remember affirmation of the
consequent?
135. If we want to validly confirm the hypothesis,we therefore test the
null, in the attempt to reject it.
Remember our double negative?
Not (Not P) = P
If we can validly conclude P, this acts as confirmation for the
hypothesis.
N.B.We never say we’ve proved the hypothesis!
Alternatively, it can be thought of as the study of consistent sets of beliefs.
Logic is probably as old as Aristotle (ancient Greek philosopher). Aristotle belonged to the well established school of sophists, which were teachers almost like tutors who had a small number of students, normally from wealthy families who were taught in public speaking and rhetoric, as well as thinking, which were all highly valued arts.
What we’ll be covering today is logic, as the discipline with which Aristotle was mainly concerned, i.e. the discipline which we might call logical argumentation
You might be thinking why do we need to do this and what have I signed up for? But learning to understand good arguments is really the bedrock of conducting and presenting good and valid research.All too often in our line of work, we are often challenging assumptions, anecdotal evidence and personal opinions, and we need to be convinced that we understand faulty reasoning and bad arguments, and that we present our own research, we’re aware of the limitations and don’t fall into obvious fallacious lines of reasoning.This particular course is about laying the foundations of this and through out a series of these training session we’ll look at particular methodologies in more detail, so you understand the purposes and limitations of various research methodologies.
POP QUIZ
In the section on quantifiers, we mentioned that Russell produced a counter-example to Father Frederick Copleston’s ‘cosmological’ argument. But what is a counter-example?
Arguments consist of some propositions, called the premises, together with another proposition, called the conclusion. For our purposes, you can think of arguments as the smallest-possible pieces of reasoning. They are the ‘atoms of reasoning.’
An argument is a piece of reasoning. It begins with some statements (propositions) called the premises, and then takes us, via a series of steps, to another statement (a single proposition) called the conclusion.
An argument is valid if, and only if, it is impossible that its premises all be true and its conclusion false.
An argument is invalid if, and only if, there is at least one interpretation of the argument under which the premises are all true and the conclusion is false.
Note, then, that arguments can’t be true or false. There’s no such thing as a ‘false argument’. Likewise, it makes no sense, on our definition, to speak of a ‘valid proposition’ (or a ‘valid belief’, a ‘valid statement’, and so on). Be careful!
So now we move to induction. Although induction was born with Aristotle, Aristotle didn’t spend much time with it and it wasn’t
If we go back to our raven example, the fact that the two ravens are black does not contradict the fact that there is a white raven in Japan. But the general rule that “all ravens are black” is inconsistent with the existence of a white raven.
A logical fallacy is one that breaks the rules of logic. You don’t actually need to learn
A non-logical fallacy is one that either be non-logical or informal, and is down to the context of the argument.
Also known as the post hoc ergo propter hoc or just propter hoc fallacy.
This is quite common in speech, as the English language has a lot of nuances of meaning
Hopefully, by this point, you’ll know what a proposition is, and understand what arguments are in the particular sense that we’ve defined them. You should also now be beginning to see what it is that makes propositions or sentences consistent, and arguments valid.
We say that we are performing a conjunction
It’s raining or it’s wet outside
It’s raining or it’s wet outsideNow, as you can see, the proposition is true when either P is true, or Q is true. Note, however, that it’s also true when P and Q are both true as well. This may strike some people as strange: when we use a sentence like ‘It’s either heads or tails’, we know it can’t be both of those things at once. There are cases, however, when that is precisely what we do mean. Consider a sentence like ‘Her grades were so good that she must have studied very hard, or else have been very, very bright.’ Nothing about this sentence says that it must be one or other of those options exclusively: a student with very good grades might, indeed, be both bright and hard-working.
This is just one of the ways in which logic is more precise than natural language. The above truth-table represents a convention of logicians to favour the so-called ‘inclusive’ form of ‘or’ over the ‘exclusive’ form. (The mathematics works out a little more neatly when we do it this way.) We could, of course, define the word ‘or’ in the exclusive sense; but for now, it’s enough to simply note the difference.
(It might be worth thinking of this connective as the slightly more verbose ‘it is not the case that…’.)
This one is a little strange. Weirdly, the only situation in which the implication is false is the one in which P is true and Q is false. In other words, the implication is false when the proposition ‘It’s raining’ is true, but the other, ‘It’s wet’ is false. (This is an impossible situation, of course; but remember that it’s the structure of the sentences we’re interested in—not what it is they actually say.)
It’s very important, however, to recognize the difference between this connective and the previous one, since a great many fallacious arguments are made by confusing the two. Consider this example from an earlier quiz: ‘If it’s raining, then it will be wet outside. It’s wet outside, so it will be raining.’ This is a perfect example of a confusion between the ‘implies’ and ‘if and only if’ connectives known as the fallacy of affirming the consequent.
In a sense, then, the truth-tables function as the foundation on which all the rest of our reasoning about propositions can be built. Nonetheless, they’re not the only helpful things we can say about the logic of propositions. We can also mention some of the basic ‘argument forms’ that logicians and scientists have long recognised to be valid. Many of these will be familiar from the exercises. Hopefully, knowing a few of them will aid you in producing good arguments later on—as well as identifying bad ones!
So common is the misuse of quantifiers, in fact, that such arguments even get their own special name: they’re known as ‘quantifier-shift fallacies’. You’ve already encountered one of the most famous examples of a quantifier-shift: it appeared on the pop-quiz at the start of this course. That argument is known as the ‘cosmological’ argument for the existence of God and it goes like this: ‘Every event has a cause. So, there must be one first cause for each subsequent event.’ In a famous debate between the Jesuit priest Father Frederick Copleston and the eminent logician and philosopher Bertrand Russell, Copleston attempted to invoke that argument in an attempt to demonstrate the existence of God. Unfortunately, Russell recognised this argument as suspect from the get-go, and produced the following counter-example: ‘Everyone has a mother. Therefore, someone must be the mother of everyone’—an obviously invalid assertion!
don’t be deterred by the fancy name (which translates roughly to ‘the way of affirming’). Modus ponens is one of the most helpful argument forms to know. Essentially, it says that, given an implication of the form ‘if P then Q’, and knowing that P, you’re welcome to infer the truth of Q on top of this.
Modus Tollens is a particularly useful argument-form in the logic of scientific experiments. For example, consider the following argument: ‘If it’s gold, then it will have atomic number 79. But it doesn’t have atomic number 79. So, it can’t be gold.’
continuing with our theme of scary names, this one translates from the original Latin to ‘reduction to absurdity’. It was introduced formally by the German polymath (and rival of Newton) Gottfried Liebniz. Reductio ad absurdum is the mother of all argument forms, and without it, modern mathematics or the sciences simply could not exist—so many of our proofs depend on it!
Galileo’s law contradicted the ages-old Aristotelian idea that bodies fall at different speeds depending on their weight. To show that this claim was incorrect, Galileo asked us to imagine dropping two objects, one heavier than the other, with both connected by a string. If we assume that heavier objects do fall faster than lighter ones, then the string would soon pull taut, as the lighter object slows down the fall of the heavier one. However, since the two objects are connected, they should fall faster than either when it is dropped separately, since their combined weight is now heavier. Thus, Galileo showed that two objects connected by a string should, on Aristotle’s view, fall faster and slower at the same time: a contradiction!
Science involves systematic observation and experimentation, inductive and deductive reasoning, and the formation and testing of hypotheses and theories
Null – there is no observed phenomenon X or Data X for Theory Y
We have observed X
Therefore, theory Y is not validPlants require light for pho..
Plants do not require light for photosynthesisWe have observed that plants do not photosynthesis in the absence of light
Therefore, the null hypothesis is incorrect