Scientific Method
Gaining the confidence
SC/NATS 1730, XXI 1
Scientific Method
How scientific
discoveries are verified
(and therefore become
“discoveries”).
The basis of confidence
in hypotheses,
supporting claims of
knowledge.
SC/NATS 1730, XXI 2
Types of Logical Reasoning
Induction:
The forming of general statements from a
number of particulars.
Deduction:
The forming of statements (assertions) based
on logical necessity.
A deduction can be a specific statement or a
general conclusion.
SC/NATS 1730, XXI 3
The empirical vs. the non-empirical
sciences
Empirical sciences
General statements
are formed from
inductions and then
used to deduce
consequences.
Example: Galileo’s Law of
All sciences of the Falling Bodies:
natural world are d = 4.9m x t 2
empirical sciences. Based upon measurements
of actual bodies falling, or
rolling, then generalized.
SC/NATS 1730, XXI 4
The empirical vs. the non-empirical
sciences
Non-empirical
sciences
Start with axioms and
deduce all
consequences.
No reference to
experience or Example: Euclid’s Proposition I.47
(The Pythagorean Theorem).
observation.
The conclusions depend only on the
Examples: logic and
axioms and the validity of the logic
mathematics. that deduced them.
SC/NATS 1730, XXI 5
The Common Sense View of
Science
Thomas Henry Huxley,
prominent 19th century
British biologist, took
the view that science is
really just a refinement
of ordinary common
sense.
Huxley made many
speeches to non-
scientists explaining (and
de-mystifying) science.
SC/NATS 1730, XXI 6
Night school
In Britain after the invention of indoor gas lighting in the
19th century, educational institutions sprang up offering
lectures and night courses for working people.
SC/NATS 1730, XXI 7
The Mechanics Institutes
The best known were
the Mechanics
Institutes, where
many educational
leaders came to give
public lectures.
SC/NATS 1730, XXI 8
Huxley at the Mechanics Institutes
At one, Huxley explained how scientific reasoning was
just common sense. His illustrative examples follow….
SC/NATS 1730, XXI 9
Huxley’s apples:
Explaining induction
Suppose, says Huxley,
that one goes to buy an
apple and is handed one
that is green. It also feels
hard. On biting into it, it
tastes sour.
After repeating the same
experience a number of
times, one might
reasonably conclude that
ALL green, hard apples
are sour.
SC/NATS 1730, XXI 10
The principle of induction
After noting several instances of
essentially the same circumstances,
always followed by the same result, we
naturally form the general conclusion that
those circumstances are always followed
by that result.
This, says Huxley, is a commonplace of
everyday life and is how we learn to live in
the world.
SC/NATS 1730, XXI 11
Induction leads to possible
deductions
The person who suffered several green,
hard apples that proved to be sour then
learns a lesson and avoids green, hard
apples in the future.
That is, armed with the induction, it can
be used as a premise in a deductive
logical argument.
SC/NATS 1730, XXI 12
The reasoning that avoids the next
sour apple
A syllogism:
Major premise:
All green and hard apples are sour.
Minor premise:
This apple before me is green and hard.
Conclusion:
This apple is sour.
This, says Huxley, is the general form of of
the scientific method.
SC/NATS 1730, XXI 13
Choosing among different
hypotheses
Preferring the probable and the consistent
When several hypotheses can each account
for the phenomena, the most probably one,
or the one most consistent with other
phenomena is to be favoured.
This is known as the principle of parsimony,
choosing the simplest explanation that covers
the evidence.
Known also as Ockham’s Razor – introduced by
William of Ockham in the 14th century.
SC/NATS 1730, XXI 14
Huxley’s homey example
On waking in the morning and coming
downstairs, one finds the teapot and
silverware missing, the window open, a dirty
hand on the window frame, footprints in the
gravel outside….
Many explanations are possible, but the evidence
points strongly to a thief. This would be the
reasonable conclusion.
In general ad hoc explanations are to be
avoided.
SC/NATS 1730, XXI 15
Ad hoc hypotheses
Ad hoc hypotheses are invented to fit the
circumstances of the particular
phenomenon to be explained. Unless they
seem probable or are consistent with
other phenomena (that appear
independent of the case at hand), such
hypotheses have little value.
It is always possible to come up with an
ad hoc explanation for any phenomenon.
SC/NATS 1730, XXI 16
Examples of ad hoc arguments
Huxley’s missing teapot and silverware:
The argument that supernatural causes were
responsible for the disappearances, e.g. that the
teapot flew out of the window on its own accord, etc.
Copernicus’ explanation of why Venus did not
show phases:
He said Venus had its own light, like the Sun.
Simplicus’ last ditch argument against the
Copernican world view:
That God could make the heavens do whatever He
pleased.
SC/NATS 1730, XXI 17
The downside of the common
sense view
While Huxley’s analysis covers many situations,
science often comes to conclusions that are very
much not common sense.
E.g., that the Earth is spinning around every day and
hurtling through space around the sun.
E.g., universal gravitation – that every body that has
mass attracts every other body that has mass with a
force proportional to the product of their masses and
inversely proportional to the square of the distance
between them.
SC/NATS 1730, XXI 18
Testing Hypotheses
When an explanatory idea about nature is
proposed, it remains a conjecture until it is
verified one way or another.
One of the key features of scientific
method is systematic testing of
hypotheses.
SC/NATS 1730, XXI 19
Case study: Puerperal Fever
A young obstetrician, Ignaz
Semmelweis, working at the
Vienna General Hospital in
1844-1848 was concerned
about the high incidence of
death from puerperal fever
in his patients and sought to
understand its cause.
SC/NATS 1730, XXI 20
Puerperal fever
Puerperal fever, also called childbed fever,
is a virulent disease that attacks women
shortly after childbirth, generally resulting
is death in a few days.
Its causes were unknown. Its incidence at
Vienna General were especially high.
SC/NATS 1730, XXI 21
General facts about pueperal fever
in the Vienna General Hospital
There were two maternity divisions, the First,
run by doctors, the Second, by midwives.
Each had students working with them.
The death rate from puerperal fever was much
higher in the First Division than in the Second.
“Street births,” women who gave birth en
route to the hospital general escaped getting
the fever.
SC/NATS 1730, XXI 22
Semmelweis sought all possible
explanations
Semmelweis looked for every possible
explanation why the fever should be
higher in his ward and sought to eliminate
them one by one.
Other than doctors versus midwives, there
were few differences in diet or general care
between the divisions.
SC/NATS 1730, XXI 23
Focusing on the differences that
there were
The differences that could be identified
included:
Priests coming to deliver the last rites to the
dying women were accompanied by an
attendant ringing a bell. In the First Division,
the priest walked through the wards to get to
the patient. In the Second Division, priests
used a side door and did not go through the
wards.
SC/NATS 1730, XXI 24
Other differences noted:
Windows in the First Division opened out to
the street. Those in the Second Division
opened into an inner hallway.
In the First Division, women delivered babies
on their backs. In the Second Division, the
turned on their sides.
SC/NATS 1730, XXI 25
A built-in control group
Semmelweis sought to eliminate possible
causes by changing practices in the First
Division to match those in the Second.
He changed the access route of the priests
delivering last rites and eliminated the bell.
He closed the windows to the outside.
He had women in the First Division deliver
babies on their sides.
SC/NATS 1730, XXI 26
Eliminating hypotheses through
modus tollens
The logical principle that Semmelweis
employed has the name modus tollens.
Modus tollens is a form of the syllogism
that demonstrates that the major premise
is inconsistent with the minor premise.
If the minor premise is known to be true, then
the major premise must be false.
SC/NATS 1730, XXI 27
Modus tollens as a tool in empirical
science
Modus tollens is the essential logical tool
to eliminate errors in empirical science.
If the major premise is an explanatory
hypothesis and the minor premise is a set of
observed facts, modus tollens can be used to
show that the hypothesis must be false and
therefore must be discarded.
SC/NATS 1730, XXI 28
Semmelweis and modus tollens
Semmelweis showed that changing the
routine of the priests made no difference
to the puerperal fever rate.
Neither did closing the windows, nor having
women deliver on their sides.
Since none of these made any difference,
these were not the causes.
SC/NATS 1730, XXI 29
The modus tollens syllogism
Call the hypothesis H.
The hypothesis will have an observable
implication, I.
Major premise:
If H is true, then so is I.
Minor premise (the observation):
I is false.
Conclusion:
H is false.
SC/NATS 1730, XXI 30
A key point
Modus tollens is only useful for eliminating
a hypothesis.
The proposed explanation H implies that the
observable fact I will be true.
If I is not true (e.g. the puerperal fever rate
did not go down), then something is wrong
with the explanation.
But if I is true, the hypothesis is not
proven.
SC/NATS 1730, XXI 31
New evidence for Semmelweis
After coming up empty handed on finding
the cause of the fever, a freak accident
gave Semmelweis a new idea.
His colleague, Kolletschka, died in a few days
after receiving a puncture wound from a
scalpel while doing an autopsy. Kolletschka
displayed symptoms similar to puerperal fever
during his brief illness.
SC/NATS 1730, XXI 32
Cadaveric matter
Semmelweis hypothesized that Kolletschka
was killed by the “cadaveric matter”
introduced into his body by the scalpel,
and that perhaps his female patients are
similarly infected by “cadaveric matter”
when being examined by medical students
who have come from doing autopsies.
Semmelweis formulates a new hypothesis
and a test for its validity.
SC/NATS 1730, XXI 33
The hypothesis and its test
implication
Hypothesis:
H = Cadaveric matter entering the bodies of
women induce puerperal fever.
Test implication:
I = If medical students wash their hands
thoroughly in a solution of chlorinated lime to
remove all traces of cadaveric matter before
examining women in the maternity ward,
incidences of puerperal fever will drop off
dramatically.
SC/NATS 1730, XXI 34
Applying the test
Semmelweis ordered medical students and doctors to use the
chlorinated lime solution when coming from the autopsy room.
SC/NATS 1730, XXI 35
Interpreting the test
The incidence of puerperal fever in the
First Division promptly fell to a rate lower
than that of the Second Division.
Eureka?
SC/NATS 1730, XXI 36
Further confirmation
Later, when his instructions were not
followed, the incidence rose again, but
was halted when washing with chlorinated
line was resumed.
Semmelweis believed he had found the
cause of the disease.
Was he justified in believing so?
SC/NATS 1730, XXI 37
The Error of Semmelweis
Semmelweis believed that cadaveric
matter (i.e., bits of corpses) was the only
cause of puerperal fever.
His reasoning:
Bits of dead bodies cause the infection.
Eliminate the cadaveric matter no infection.
SC/NATS 1730, XXI 38
A troubling unexpected case
A woman had been admitted with cervical
cancer and had been placed in the
maternity ward.
She had been examined by the doctors
and students, who then went on to
examine the other women in the ward,
without washing their hands.
All the other women in the ward
developed puerperal fever.
SC/NATS 1730, XXI 39
The hypothesis, H, was too
restrictive
Semmelweis had believed that only matter
from corpses conveyed the infection. He
had not considered that the problem was
putrefaction.
There was no theory of microbes at the time.
Disease was not understood to be caused by
bacterial infection, since bacteria were
basically unknown.
SC/NATS 1730, XXI 40
The Fallacy of Affirming the
Consequent
Semmelweis had unwittingly committed a
logical fallacy, known as the fallacy of
affirming the consequent.
The form of the fallacy:
If H is true, then so is I.
I is true.
False conclusion: H is true.
SC/NATS 1730, XXI 41
Semmelweis’ fallacy
His implication, I , was that washing the
hands after doing autopsies will prevent
the fever.
His hypothesis, H, was that cadaveric
matter was the sole cause of the fever.
But the reasoning is fallacious because I
can be true when H is false.
E.g., apples that are not green and hard can
also be sour.
SC/NATS 1730, XXI 42
Falsification
It is an inescapable feature of empirical
science that a hypothesis, or a theory, can
never be fully verified as true.
It is possible to show that a hypothesis is
false (using modus tollens), but not to be
true.
SC/NATS 1730, XXI 43
Confirmation
The best that can be done is to confirm that a
hypothesis is consistent with other hypotheses
and theories, and has many true implications,
and therefore, probably, is true as far as we
know.
The logical form of confirmation:
If H is true, then so are I1, I2, I3, …, In.
Evidence shows that I1, I2, I3, …, In are all true.
Conclusion: H is probably true.
SC/NATS 1730, XXI 44
