The Scientific Method Reflections From A Practitioner

THE SCIENTIFIC METHOD
THE SCIENTIFIC
METHOD
Reflections from a Practitioner
Massimiliano Di Ventra
University of California
San Diego, USA
1
3
Great Clarendon Street, Oxford, OX2 6DP,
United Kingdom
Oxford University Press is a department of the University of Oxford.
It furthers the University’s objective of excellence in research, scholarship,
and education by publishing worldwide. Oxford is a registered trade mark of
Oxford University Press in the UK and in certain other countries
© Massimiliano Di Ventra 2018
The moral rights of the author have been asserted
First Edition published in 2018
Impression: 1
All rights reserved. No part of this publication may be reproduced, stored in
a retrieval system, or transmitted, in any form or by any means, without the
prior permission in writing of Oxford University Press, or as expressly permitted
by law, by licence or under terms agreed with the appropriate reprographics
rights organization. Enquiries concerning reproduction outside the scope of the
above should be sent to the Rights Department, Oxford University Press, at the
address above
You must not circulate this work in any other form
and you must impose this same condition on any acquirer
Published in the United States of America by Oxford University press
198 Madison Avenue, New York, NY 10016, United States of America
British Library Cataloguing in Publication Data
Data available
Library of Congress Control Number: 2018931028
ISBN 978-0-19-882562-3
DOI 10.1093/oso/9780198825623.001.0001
Printed and bound by
CPI Group (UK) Ltd, Croydon, CR0 4YY
To
Elena, Matteo, and Francesca
in gratitude for their central and unique role in my life
Preface
I am a theoretical physicist who has been working for

over twenty years in academia, publishing regularly in
refereed journals, teaching graduate and undergraduate
classes, training several graduate students and post-
graduate researchers, traveling the world to exchange ideas
at conferences, and enjoying the creative aspect of my
profession.
Over the years, however, while discussing with educated
laypeople interested in Science, students, and even some
researchers, I have come to realize that they attribute to
Science the ability to answer questions that have always
been addressed by other sources of knowledge, in particu-
lar, Philosophy or Religion.
With this in mind, I felt the need to discuss these issues
with my own students. In particular, I wanted to better
viii Preface
convey to them both the reach and, most importantly, the

limits of Science and the knowledge we can acquire from this
human enterprise.
Encouraged by their feedback, I offered to teach a class on
the scientific method in the Department of Physics of the
University of California, San Diego, in the fall of 2012 and
spring of 2013.
While researching for my lectures, I realized that the
misjudged role of Science is often amplified by the frequent
mis-representation of scientific discoveries or theories by the
media, with us scientists sometimes complicit in this act.
The unfortunate result is that oftentimes—pressed by
the obvious desire to increase readership—a scientific
hypothesis is elevated to a fact. These “news” spark the
imagination of laypeople, sedimenting into our collective
consciousness “truths” that have much more in common
with science fiction than reality.
Navigating scientific topics may then seem daunting. Not
all of us are grounded in advanced Mathematics or Physics
or other disciplines necessary to judge the value of a scien-
tific statement.
Even the practitioners of a particular discipline cannot
claim expertise in all the other subjects of scientific
endeavor, and hence cannot discuss with familiarity all the
body of scientific knowledge that is continually generated,
seemingly at an ever-increasing rate.
While this working knowledge may not be available to all
of us, it is my conviction that it is unnecessary to evaluate
a scientific claim. In fact, Science rests on a well-defined
methodology that provides a guiding set of rules for learning
Preface ix
about Nature, and, at the same time, defines the boundaries

of scientific inquiry.
Whether we discuss Biology or Chemistry or Physics, the
scientific method is common to all Natural Sciences and is the
fabric and the backbone of scientific knowledge. Familiarity
with the scientific method is not only a necessary tool in
modern times, but it is also well within reach of everyone,
whether a practitioner like me, or an interested layman.
This short book at a level accessible to first-year under-
graduate students of all Natural Sciences—and quite pos-
sibly to high school students—attempts to provide such a
tool to a wide audience. It originates from the notes I have
written for my class on the same subject.
By Natural Sciences I mean all those disciplines that
describe phenomena occurring in Nature, such as Physics,
Chemistry, and Biology. (Some of the points discussed in
this book may also be relevant to the Social Sciences, such as
Economics or Psychology, when empirical approaches are
employed.)
Although Mathematics is the language in which we
express many of these phenomena, its method and goals
are substantially different and will not be treated in this
book.
The book addresses the following concepts: objective
reality and approximate description of natural phenomena;
the role of the observer; the difference among objective
facts, hypotheses, and theory; the meaning of “falsifi-
ability”; the “absence of democracy” in the pursuit of
scientific truths; and the fundamental and inalienable role
of experimental evidence in scientific knowledge.
x Preface
I intend it to be an easy read, and therefore I have made

an effort to avoid difficult concepts. The examples I have
chosen to clarify the method are taken from well-known
and easy-to-follow facts so that the reader should easily
focus on the method itself rather than the validity of a fact.
My hope is that starting from such examples, by reflect-
ing on the general method, the reader will be able to criti-
cally sort through other types of scientific claims, and judge
for himself whether they stand on a solid methodology of
study or on shaky grounds.
It is, in fact, my belief that a universal understanding of
what the methodology of a scientific study is, its reach and
limitations, would strengthen our culture, better inform
our decisions in related matters, and channel our creativity
in building a better future.
Although this is a book on the method of scientific
inquiry, I do not intend it to be a treatise on epistemology—
“the theory of knowledge.” Rather, it is a collection of
personal reflections on scientific methodology itself as
experienced and used daily by a practitioner.
In this respect, my discussion and understanding of the
limits and possibilities of Natural Sciences are very much in
line with those of a fellow physicist of the 19th century who
wrote extensively about the structure of a physical theory:
Pierre Duhem. The famous quote “to save the phenom-
ena,” first attributed to Plato, and espoused and extended
in Duhem’s work, summarizes, in my view, the essence of
what a scientific theory should aim for.
Despite this not being a book on epistemology, it will start
from, and build upon the philosophical/logical foundations
of Science. This starting point is nowadays more important
Preface xi
than ever. In fact, by forgetting its foundations, we trans-

form Science into some form of bad “religion” that we may
call “Scientism.”
Scientism is very detrimental to Science per se and should
be rejected unequivocally by all, laypeople and practition-
ers alike.
The addition of a chapter on this form of “religion,”
and one on the “What” and “Why” questions, is then my
humble attempt to clarify and rediscover the foundations
of what it truly means to do Science in an age when the
noise of uninterrupted news and fads has a negative impact
on our collective knowledge.
My hope is that this book will help rekindle a much-
needed interest in the centuries-old structure and founda-
tions of this marvelous human enterprise we call Science.
La Jolla, 2017
PS: Why the cartoons? You have heard that “a picture

is worth a thousand words.” I personally think that a
humorous picture renders the reading of the present
subject lighter. It also conveys the message better and
expresses to the reader that we can always discuss serious
arguments with a joyful demeanor.
PPS: The book is illustrated by MD2 (a.k.a. Matteo Di
Ventra).
PPPS: “No one is more hated than he who speaks the Truth.”—Plato.
· · · ∞ · · · PS: “My desire is to befriend everybody, but my biggest
desire still is to befriend Truth.”—my own rendering of Aristo-
tle’s famous quote on Plato.
Contents
1. Science without Philosophy? 1

2. Material World and Objective Reality 6
3. First Principles and Logic 14
4. Natural Phenomena and the Primacy
of Experiment 24
5. Observation and Experimentation 33
6. The Role of Human Faith in Science 36
7. Approximate and Limited Description
of Natural Phenomena 42
8. Hypothesis 45
9. Theory 53
10. Competing Theories 62
11. Can One Theory Be “Derived” from Another? 66
12. Verifying or Falsifying? And What? 69
13. Don’t Be a Masochist! 75
14. “Consensus” in Science? What Is That? 83
15. Flow Chart of the Scientific Method 88
16. The “What” and “Why” Questions 89
17. “Scientism”: Abusing the Scientific Method 96
18. Final Thoughts 103
Further Readings 111

About the Author 112
Index 113
1
Science without Philosophy?
A first question to settle before embarking further into the

topic of the scientific method is the following.
Can Science stand on its own feet without the “Love of
Wisdom,” a.k.a. Philosophy?
In an age in which Science has conquered the podium
as one of, if not the most respected human endeavors,
we have been led to believe that we may avoid any
philosophical argument to define its method and make
it a solid enterprise.
So, we hear statements of the sort “Science is the only way
we know truth”1 or “Philosophy is dead.”2
1 A simple Google search would show that this statement has been
written in too many articles to cite. Emphasis is mine.
2 S. Hawking and L. Mlodinow, The Grand Design: New Answers to the
Ultimate Questions of Life (Bantam, 2010).
The Scientific Method. Massimiliano Di Ventra, Oxford University Press (2018).

© Massimiliano Di Ventra.
DOI: 10.1093/oso/9780198825623.001.0001
2 The Scientific Method
Let us, for instance, analyze the first statement: “Science

is the only way we know truth.” This statement begs the
following question: “How can Science prove this statement
to be true?”
In no way could we ever perform an experiment that
would validate it!
Therefore, the above statement is not scientific in itself. It
refutes itself, and therefore it is not true.
But then, if empirical tools are the only ones at our dis-
posal, how did I arrive at the conclusion that the statement
is not true if I cannot prove it scientifically?
I came to that conclusion by means of what we humans
innately possess, and which has been formalized from at
least the very first philosophers, such as Aristotle: logical
thinking.
The Logic I am talking about here is the foundation of
Philosophy and precedes Logic as a mathematical discipline.
It is this type of “philosophical” Logic upon which we build
any knowledge, whether scientific or otherwise, and we use
it every day, often without even realizing it consciously.
For instance, suppose we go to work following a particu-
lar road. That road is always dry, but one day we observe it
to be wet.
We would then conclude, without actually having a
direct proof, that something (e.g., water or some other
liquid) must have rendered the road wet. The road cannot
become wet if left to its own devices.
Somehow, we know that some other agent, other than
the road itself, must have caused the road being wet, even
though we do not know what that agent is.
Science without Philosophy? 3
We reach that conclusion by implicitly using a form of

logical thinking built upon our experience.
If someone asked us how we arrived at such a conclusion,
we would simply say that it is logical to think that if the
road is now wet, when it wasn’t the other days I traveled
it, something must have made it change its state of dryness.
It is just common sense to reach that conclusion.
The reasoning that goes from the direct observation (the
wet road) to claiming that an agent must have caused it to
be wet is then simply “logical.”
If a friend would tell you that, no, the fact that the
road is now wet does not mean anything, namely that its
“wetness” is caused by absolutely nothing at all—it just is—you
certainly would look at him bewildered!
As it is for any reasoning we make in our daily life, we
will see that also Science does need the logical, founda-
tional aspects of Philosophy to even define its tenets and
method. Without such a logical foundation, nothing we do
in Science would make any sense.
On the other hand, not all logical statements we make are
scientific even though they pertain to material things, like
the “wet road” example.
In that example, both the cause (water or some other
liquid) and the effect (the property of the road being wet)
are natural. I am not discussing any “supernatural” effects (a
ghost showered it?). And, of course, I used my senses (my
eyesight or touch) and previous experiences to reach a valid,
logical conclusion.
However, despite all this, I never invoked or used the
scientific methodology that I will discuss in this book.
I wanted to stress this point, because, in conversations

and journal or news articles, we do sometimes encounter
examples (and I will provide some in the final chapters) of
what appear to be “scientific statements” that may be logical
per se, but are not scientific at all!
These are the statements we need to guard ourselves
against, when we go about doing Science or judging the
scientific content of some “theories.”
Finally, even if we do not discuss material objects, we can
still use logical thinking to validate or invalidate a given
statement, namely to know its truth value.
For instance, the reasoning that I used to invalidate
the statement I discussed at the beginning, although less
obvious than the “wet road” example, and not related to
our daily experiences, still falls into the category of logical
thinking.
In fact, as I will discuss later in more detail, since Science
can only deal with materially testable, experimentally ver-
ifiable statements, the quote “Science is the only way we
know truth” is logically wrong, because its truth value cannot
be tested experimentally.
And if someone would tell you otherwise, you may also
look at them bewildered, because they are behaving, in this
case, “illogically.”
So, instead of concluding that “Philosophy is dead,” we
must acknowledge that it is very much alive. As human
beings we (should) apply our rational thinking (common
sense) in any enterprise we engage in. Science is no
exception.
I will only discuss those aspects of Logic that are critical to
the scientific method. Before doing this, however, we need
to define what the object of scientific study is.
Science without Philosophy? 5
Takeaways from this chapter

• Logic is very much essential to do Science and is at the
core of its method.
• Statements we make can be logical and yet not scientific.
• Science is not the only way we know truth. Other intel-
lectual domains, such as Philosophy and Religion, are
legitimate sources of knowledge.
• Philosophy is definitely not dead.
2
Material World and Objective Reality
All scientific disciplines tend to specialize in particular

aspects. Biology studies living organisms, Physics the
temporal and spatial relations between inanimate objects,
and so on.
However, irrespective of the particular subject, all Nat-
ural Sciences deal with objects and events that we observe
either directly through our senses, such as sight, smell, or
touch, or indirectly through their extensions, namely instru-
ments that we build to explore objects and events whenever
we cannot probe them directly with our senses.
For instance, certain stars or even planets in our solar
system are difficult, or outright impossible to see with our

DOI: 10.1093/oso/9780198825623.001.0001
Material World and Objective Reality 7
naked eyes. Telescopes have been built to precisely expand

the reach of our eyes. In general, any instrument we have
built to explore our surroundings expands the reach of our
human senses.
Irrespective of the subject of a particular scientific disci-
pline we can then safely say that the object of study of the
Natural Sciences is the material world.
In particular, we study the relation between material
objects that surround us through our interaction with them.
What we observe is a series of phenomena, namely what
happens to material objects (whether inanimate or living)
when they are in interaction with each other and, to some
degree, in interaction with the instruments that we use to
probe them.
We call all this material reality or material world.
Here, we can immediately make an aside and say that
Science is only concerned with the material world. It cannot
and should not venture into anything that does not pertain to
the material world.
As we will see in a moment, the tools of Science are
empirical; the only information that we can extract from
the material reality comes from our interaction with our
surroundings.
By “interaction” I mean any direct or indirect action
aimed at exploring Nature via our instruments, and by
“surroundings” anything that goes from the infinitesimally
small (the sub-atomic world) to the extremely large (the
Universe we observe).
There may well be a super-natural reality. However, the
word itself, “supernatural,” puts it outside (“supra” from
Latin means “above”) the natural (material) reality Science
can address. Of course, we can, and do deal with it within

the confines of Philosophy or Religion, but not Science.
For the sake of Science, we must then limit its reach to
the material reality and nothing else.
This is, however, not enough to establish that such a
reality is the same whether I observe it, or some other person
observes it.
Scientists implicitly take for granted that this is the case.
For instance, if I and another scientist observe the Moon
under the same or similar conditions, we would observe
the same object, possibly displaced in space because of its
relative motion with respect to Earth, nonetheless still the
same object.
In other words, what should we expect of the outcome of
several observations of the same phenomenon by different
observers but under similar conditions?
Answer: the same result or data, at most corrected for
the conditions under which the experiment has been per-
formed. Yesterday, the Moon was in a different position
than today, but today’s position of the Moon could have
been predicted yesterday! So, we agree that the object of study,
the Moon, has not changed, only its relative position with
respect to the observer.
It is then imperative for Science to declare the following:
Object of study
The unique material reality which is objective—namely independent
of the observer—and which is accessible via our senses or their
extensions.
I have encountered people that would deny the existence

of such an objective reality. Although this is often the
subject of movies or some “beliefs,” I cannot truly fathom
a scientist claiming the same.
Denying the existence of an objective reality that is
present irrespective of the observer is equivalent to denying
the very existence of Science and its method.
In order to stress how problematic such a denial is for Sci-
ence, imagine if a scientist were to perform an experiment
under certain conditions and observe a “reality” different
than that observed by another scientist who performs the
same experiment under the same conditions.
Say, one looks into a microscope and observes bacteria,
and immediately after another scientist takes his place and
observes viruses with the same microscope. These two sci-
entists would come together and exchange the observed
data, and find them all different.
Which ones are true? How can they build any under-
standing of what is being observed?
If that were the case, they would not even be able to agree
on what the object of study itself was!
As I will explain later, in a case like the one discussed
above, the scientific method would suggest that one of the
two, or both, must have made a mistake in their observa-
tions, or accidentally replaced the object to observe. Ulti-
mately, all these experiments would have to be repeated
until they all agree on the data being collected!
Note that this is also true for quantum phenomena.
In that case, an individual measurement of a given physical
property may be possibly different for different observers.
However, the collective outcome of all measurements on the
same physical property must be the same irrespective of the

observer, if the system is probed under the same conditions.
Physicists would then extract from those measurements
important parameters, such as the average value of the
given quantity under study, or other physical parameters
that characterize the phenomenon they are trying to
understand. Those parameters are ultimately the ones on
which different observers have to agree.
If some substantial discrepancy between the outcomes of
measurements performed by, say, two different teams of
observers occurs, then something must have gone wrong
in one or both experiments.
It would be foolish to think the researchers could con-
clude that they live in different realities, hence “all is good.”
If that were the case, the whole edifice of Science would
come falling down, and nothing we do would make any
sense!
Note that I am not saying all of them must agree on the
interpretation of the data, only on the data themselves. We can
debate whether viruses are alive, but we should recognize
that bacteria are bacteria, not viruses.
I will get back to the role of the experiment more in depth
later because it is of fundamental importance. For now, it
suffices to agree on the existence of an objective reality.
Now that we have established the existence of such a
unique material reality, I am ready to define a point that
I have already anticipated.
It is a fundamental tenet of the Natural Sciences. Even
more, it is an operational tenet, since it defines, at the outset, how
we should go about doing Science itself. This tenet is its limit
of inquiry:
Limit of inquiry
The only object of study of the Natural Sciences is the mate-
rial reality. Nothing other than the material reality should be
approached by the inquiries of the Natural Sciences, because
nothing other than the material world can be probed experimentally
by our senses or extensions of our senses.
Note that this limit of inquiry is trespassed by the previous

(false) statement that “Science is the only way we know
truth.”
The limit of inquiry puts a boundary on what Science
can and cannot access, and declares that the only truths that
Science has any right to discuss are those that can be probed
experimentally, namely only those pertaining to the mate-
rial world.
In other words, there is no point in invoking Natural
Sciences to prove or disprove the existence of God, the value
of freedom, the meaning of beauty, and anything else that
is beyond our experimental reach.
Mind that these are all valid, legitimate, and fundamental
questions, without a doubt worth exploring, but with tools
that belong to Religion or Philosophy, not experiments, and
as such they do not belong to the realm of study of the Natural
Sciences.
As I have already anticipated, this is also true for any
statement made by scientists: if it cannot be tested experimentally
it is, at best, a mere hypothesis, or personal opinion!
As I said, these statements could even be logical and yet
not scientific. To stress this difference, let us go back to the
example of the “wet road.”
In that example I concluded that if I found a road wet,

while it was dry before, then water or some other liquid must
be responsible for its new state of wetness. The road (the
material making it) does not have the attributes of being
wet if left to its own devices.
So far, I simply reached a valid conclusion using logical
thinking. What would it take to make it a scientific argu-
ment, namely, one that is testable by experiments?
Never did I ask what phenomenon has caused the liquid I
observed on the road to be there. How did it occur that the
liquid happened to be there at that particular point in space
(the road) and at that time (the day I observed it)?
In summary, if I ask “How come water was present on
that particular road on that particular day?” I am actually
asking a very specific and completely different question.
In this case, I am really looking for the natural phenomenon
that made water fall onto the road and rendered it wet. The
question can be addressed in its entirety within the scientific
method.
I could say (hypothesize) that it rained the day before I
traveled that road and, hence, rain is the culprit. I can
indeed test this hypothesis by looking at the weather record
(if it exists) in the area of the road.
If the weather record confirms that it rained that day (at
least with some likelihood), then I can conclude (with some
level of confidence) that my statement (“the rain caused
water to fall on that road on that day”) is consistent with the
observation that the road was indeed wet.
Of course, I still used logical thinking in the above
example (e.g., in the highlighted “If . . . then” statement),
but I went a step further by making statements on the
possible phenomenon that could have produced the observed

wetness.
And I have done this all along by working only within the
limits of inquiry of the Natural Sciences, namely I was only
looking for material causes, and not some super-natural ones.

• We access a unique material reality via our experimental
probes.
• The fact that this reality is unique is a necessary, logical
premise of doing Science in an objective way, namely
independently of the observer.
• Science can only deal with this objective material
reality.
• Science cannot deal with subjects or questions that do
not lend themselves to experimental tests.
• All scientific statements need to be logical, but not all
logical statements need to be scientific.
• We can reach valid logical conclusions without ever
employing the scientific method, namely without
ever needing experimental tests of such conclusions.
3
First Principles and Logic
Once we have settled on the issue that there is such a thing

as an objective reality—the only entity that we can explore
scientifically—we can now ask the question “In which way
do we discover it?”
Unlike animals, we are rational beings and we can let our
imagination roam free. However, when it comes down to
Science we have to recognize that we have quite a few, well-
defined, and non-negotiable boundaries that need to be satis-
fied. I call “boundaries” the constraints of logical thinking
I mentioned before.
We proceed, namely we reason, with a well-defined Logic
that is built in our very nature. This Logic is based on

DOI: 10.1093/oso/9780198825623.001.0001
First Principles and Logic 15
several important first principles or self-evident truths (or laws of

thought).
The statements I will write down are self-evident because
they cannot be proved neither mathematically nor philo-
sophically, nor using Logic itself. In fact, they are its foun-
dations.
However, if I tell them to you, you would certainly recog-
nize them to be true without any doubt or second thought.
They just are! In an informal way, they summarize the core
of what we call common sense.
Anybody who doubts these principles is very limited, if
not unable to discuss rationally any argument! Worse than
that, try to deny any one of these truths and see if you
can even make sense of your own thinking, let alone of a
discussion with someone else.
These are the three foundational laws of thought whose
formulation is typically attributed first to Aristotle.1
The law of identity

An object is the same as itself.
The law of non-contradiction

One cannot say of something that it is and that it is not in the
same respect and at the same time.
The law of the excluded middle

For any proposition, either that proposition is true or its negation
is true.
1 Metaphysics 4.4, W.D. Ross (trans.), GBWW 8, 525–6.

The attentive reader would immediately recognize that

I have used them in my previous discussion without men-
tioning them explicitly. For instance, when I said that any
one of those bacteria observed is a bacterium—they are
what they are—I made use of the “law of identity.”
The “law of non-contradiction” is the statement that truth
cannot be self-contradictory because something cannot be both
true and false at the same time and in the same way.
For instance, bacteria cannot be bacteria and viruses at the
same time and in the same way. Or if a scientist says he
observes a cat, either that cat is alive or its negation is true
(“the cat is dead”). He cannot claim the cat is both dead and
alive at the same time and in the same way.2
You should have also immediately recognized that the
sentence I wrote before, “Science is the only way we know
truth,” is false since its truth value (whether it is true or
false) has no means of being proven only from its own
meaning (namely that only Science is able to find truths).
In other words, that statement is a contradiction because
it excludes the possibility of another one (“there is some
other, non-scientific way of knowing the truth”), and yet it
claims to be true: if there is no other way of knowing the
truth other than Science, how do I know this statement is
true, since the statement itself is not a scientific one and yet
claimed to be true?
2
Quantum Mechanics does not contradict this law (or any other law
of thought for that matter), as some news articles may seem to imply.
The famous Schrödinger’s cat gedanken experiment acknowledges that
when the cat is observed it is either dead or alive, not in a superposition
of dead and alive states. I’ll come back to Quantum Mechanics and its
apparent issues with these laws of thought sometime later.
This is, indeed, an application of the “law of excluded

middle,” which says either that a statement is true or its
negation is true.
To these three first principles we still need to add another
one, which, in fact, I have already implicitly used.
Suppose I see a ball rolling on the floor but I do not see
what made it move. Even if I do not see what caused such
movement, I know for sure that the cause of such an effect
(the ball rolling) must have been something or someone
pushing it somehow.
In other words, we can stipulate the law of cause and
effect:
The law of cause and effect

For an event to occur at a given time, there must be another event
that has caused it.
There are actually two ways to look at this principle: one

philosophical, one scientific.
In Philosophy we can abstract the two entities of this law
(cause and effect) and consider logical conclusions that have
nothing to do with Science.
As an example, one could start from the observation of
motion of an object (e.g., the ball rolling) and conclude that
some cause must have generated such motion (e.g., a child
kicking the ball). But, in turn, that particular cause can be
viewed as the effect of some other cause, and so on.
Strictly speaking, in Philosophy, a “causal regression”
(whether the one I discussed above or any other) does
not need to be confined to a specific set of objects or
phenomena. In fact, those causes could be anything in the

Universe, and affect anything, ultimately even any notion,
e.g., space and time, or our own thoughts.
An important consequence of all this, which we may
want to mention now, pertains to the philosophical mean-
ing of “creation.”
In fact, if we abstract this “causal regression” to anything,
either it may never end or we can subscribe to the notion
that there exists a “First Mover,” who is the ultimate cause and
is “self-sufficient,” namely does not need to be caused.
We typically call this “First Mover,” God. He would have
put into existence (hence in motion) the whole reality, including
space, time, matter, etc., from the true nothing, the philosoph-
ical concept of absence of anything that exists.
Philosophically, this is creation ex nihilo, precisely because
it comes from the absence of anything.
On the other hand, let’s not confuse this “absence of
anything” with the “vacuum” that is discussed in Quantum
(Field) Theory. The quantum vacuum is a physical entity
with physical attributes of space, time, energy, etc.
When we say “absence of anything” in Philosophy, we
really mean everything, including absence of those attributes!
Note, however, that although I have used logical think-
ing to discuss the presence or absence of a “first cause,” the
conclusions I have reached are within the realm of Logic (or
Philosophy in general), but are definitely not scientific.
Why? Because, while I may logically reach the conclusion
that there exists a supernatural being (or even a “supernat-
ural reality”), outside even space and time, the very fact that
it is supernatural puts it outside of the limit of inquiry of the
Natural Sciences.
In simple terms: there is no way to experimentally test, with

any possible physical instrument, such a reality, even if it
existed.
Now that we have briefly discussed the reach of the law
of cause and effect in a philosophical context, let’s move on
to its meaning in Science.
In a scientific sense, we use the law of cause and effect
(causality) in a much more restrictive way.
Rather than considering its “abstract” or philosophical
conclusions, we employ it in a more practical way.
So, what qualifies as “cause” of a given event, in a scien-
tific sense, and when should we be content that we have
found it?
This is a very important point to consider since it defines
the boundaries of what we can say about a given phe-
nomenon or a set of phenomena.
In fact, in Science, answering the “cause question” of a
phenomenon is tantamount to deciding how we approach
it, in terms of experiments to perform, and the theoretical
description we provide of such a phenomenon.
In other words, the scientific “cause” of an effect is related
to our ability to measure (test) that effect; for example, how
far our measuring instruments can go in observing the nat-
ural world, as well as the level of description we, scientists,
are willing (or able) to provide of such a phenomenon.
To make things clearer take as example the rolling ball.
I may discover that it was kicked by a child, and I could be
satisfied with this plausible cause (“explanation”).
However, I could be more descriptive and say that the
child’s foot exerted a force and imparted a change in
momentum on the ball, which then moved.
Yet more: the atoms of the child’s foot, collectively,

through their mutual interaction, received energy from
the rest of the leg and moved at a given speed. These atoms
interacted with the atoms making up the ball, and through
repulsive interactions they transferred momentum to
them. Or the quarks of the nuclei of the atoms. . . . I think
you get the point.
This exercise needs to stop at some point. At which point
it stops is determined by both the investigator himself and,
to a certain degree, the instrument of investigation. The
investigator may not be satisfied with a particular descrip-
tion of an effect and may then advance a new hypothe-
sis of what may have caused a particular effect, but the
instruments at his disposal to test such hypothesis are not
available.
Following the above example, we could say that the
ball moving is the result of the atoms in the leg of the
child, but our naked eyes have no means of testing such a
statement. Therefore, if our eyes are the only observational
instrument at our disposal, we can be content with the
description (“explanation”) that the child’s leg has imparted
momentum to the ball, without worrying about the atoms
in his leg.
Of course, this does not preclude a researcher from
making predictions about a more “elementary cause” of that
particular phenomenon, even though the instruments
at our disposal do not allow a test of such a prediction
as of yet.
Therefore, even if the eyes are the only instruments at
our disposal, the researcher could still argue that there is a
more “elementary cause” behind the momentum imparted

by the child’s foot, and posit the existence of atomic entities
in the child’s leg.
This researcher’s description of what caused the phe-
nomenon would then be considered more “refined” or
“elementary.”
However, as I will explain in depth later, to qualify as
scientific, this description must lead to testable predictions that
could come to fruition—read, be tested experimentally—when
new instruments and experiments have been devised pre-
cisely to test such predictions.
Otherwise, this description will simply remain in the
realm of hypothesis, or worse, be fruit of the researcher’s
imagination.
In conclusion, in Science the level of description of an
event and its “cause” are not abstract concepts unrelated to
the observer(s). They are determined by the observers’
degree to which they approach that particular natural
phenomenon and the experimental probes at their disposal.
Note that, due to the limit of inquiry I discussed earlier,
it follows that the cause of a given effect must be natural.
As scientists, we are not allowed to invoke “supernatural”
causes—a ghost, a miracle—namely causes that are outside
the material reality , which is the subject of study of Natural
Sciences.
We may, in addition, clarify that causality is distinct from
correlation. The latter simply establishes a relation between
two events. For example, when I saw the ball rolling, I also
heard a loud noise. The two events may be correlated, but
not necessarily one follows from the other.
Hopefully, all this discussion should have elucidated even

further the difference between scientific statements and logical
ones.
Even though both need to build upon the foundations
of Logic, the latter ones do not need to correspond to a
scientifically testable material reality.3
I will exemplify these issues further in the book’s last
chapters because very often they create confusion on what
Science can and cannot address. For now, let us move on
and stress even more the fundamental role of experiments.

• We use the foundational laws of Logic (first principles)
to think rationally and do Science.
• The application of the law of cause and effect is some-
what different in Philosophy from that in Science.
• The philosophical “vacuum” is different from the
physical “vacuum.” The former is absence of anything,
including physical attributes. It cannot be measured. The
latter has physical properties that are measurable.
• The scientific “cause” of an effect is related to both our
ability to measure the phenomenon we are interested
in and the level of description we decide to use at the
outset.
3 The majority of—and in a very strict, formal sense, all—statements
in Mathematics are like this. Mathematical statements only need to

follow logically and unequivocally from the foundational axioms of
Mathematics, regardless of whether they correspond to a material
reality.
• This description, however, needs to lead to testable pre-

dictions to be considered a valid, scientific description.
• Otherwise, it remains within the realm of hypothesis
or simply fruit of our imagination.
4
Natural Phenomena and the
Primacy of Experiment
We have now established that the object of study of the

Natural Sciences is the (unique) material reality around us.
We, the observers, are able to explore such a reality with
our senses or extensions of our senses by probing the inter-
actions between the material objects and our instruments,
interactions that we call natural phenomena.
For instance, when I look at two objects colliding, my
eyes collect the light that bounces off those objects for the
whole duration of my observation. My brain then analyzes
that information to make sense of what I see. Without the
interaction of that light with my eyes, I would not be able
to see the objects at all. The same goes for any other type of
measurement we make of phenomena.

DOI: 10.1093/oso/9780198825623.001.0001
Natural Phenomena and the Primacy of Experiment 25
However, what part of the material reality do we actually

explore?
First of all, we need to decide what type of phenomenon
we are interested in.
Maybe we are interested in the motion of the planets
in our solar system. Or we are interested in how a cer-
tain organism interacts with its immediate environment or
some other organisms.
Irrespective, our goal is not to study the entire material real-
ity at once. Indeed, that would be an impossible proposition.
We are limited beings and our instruments have a limited
reach.
Therefore, we limit our inquiries to an extremely small
subset of such a reality. In fact, as we discussed in the
preceding chapter, we are limited by the instruments we
use to probe it, since the latter ones determine not just the
type of phenomena we can address, but also, for a given
phenomenon, the type of information we can extract from
our measurements.
We can also reverse this argument: as scientists, the only
aspects of the material reality we can explore are those
within reach of our experiments. To be precise, we can only
explore what our senses and extension of our senses (our
measurement apparati) allow us to do.
Of course, by inventing new instruments we expand the
reach of possible phenomena we can study. Hence, we can
interrogate more of the material world.
However, anything that does not result from our mea-
surements of the material world is either a hypothesis or
pure speculation.
Despite what we may sometimes hear, unless we are

able to experimentally investigate the world around us, we, as
scientists, can only offer hypothesis or opinions (“educated
guesses”?), which cannot contribute to the body of knowl-
edge we call Science.
This shows why experiments constitute both the starting
point and the end goal of the Natural Sciences.
Now that we have realized that we can probe, at any given
time, only a limited aspect of the material world with our
instruments, we can discuss how we go about doing that.
Ideally, we would like to keep the interaction of the
measurement apparatus with the material objects we want
to study as small as possible so that the apparatus does not
perturb considerably what is happening to those objects.
Of course, this interaction cannot always be made small,
but at least we should be able to know the extent of its
perturbation on what we observe.
If we know how much the probes influence the measure-
ment, by repeating the experiment many times, even with
different types of probes, we should be able to reproduce the
same results within the measurement errors.
For instance, if I want to measure the distance between
me and an object, I can measure such a distance directly
using a ruler, or, if that is impractical because the distance
is quite large, by sending radio waves (using a radar) that
bounce off the object and return to my position after some
elapsed time. Through the knowledge of the speed of radio
waves in air I can then determine such distance.
Regardless of the measurement process used, to claim
that such a distance is, say, a kilometer, I need to repeat
such a measurement many times. Each measurement will
be slightly different from the others (say, one kilometer

plus or minus a few meters) due to factors that are difficult
to control. (For example, in the case of the radio waves,
possible perturbations from the air density variations in
between me and the object.)
However, if I repeat the measurement many times my
confidence in the fact that the distance is one kilometer
increases, and I can ultimately assign that value averaged over
all the measurements, plus or minus some of its fraction
(the measurement error or uncertainty).
If I were able to measure such a distance with a ruler,
then such type of measurement would again need to be
repeated many times, producing an average value of the
distance and the associated error. By comparing the two
types of measurements (one with the ruler, the other with
the radar) I can say that the two are consistent (or that “they
agree”) if the average of one and the average of the other are
within errors from each other. For instance, the measure-
ment with the radar gave me a distance of (1000 ± 1) m,
while the ruler one produced (999 ± 2) m. Then the two
measurements are consistent.
If, instead, the ruler measurement gave me (990 ± 2)
m then, even if I take into account the error in this mea-
surement, its maximum value would be 992 m, while the
radar measurement would be, at a minimum, 999 m which
is still 7 m larger than the previous maximum value. The
two measurements are inconsistent (they “do not agree”) within
their respective errors.
If the two measurements are inconsistent, we may per-
form a third type of measurement with different instru-
ments (e.g., by walking all the way with equal steps; tough
but doable), or we realize that one of the two (or both)

had, in fact, some systematic error that we did not detect at
the outset. For instance, the measurement with the ruler
was skewed by our inability to accurately reset its position
at every step of the measurement, thus giving an effective
larger error than previously estimated.
Regardless, a given phenomenon is declared to be rep-
resented by objective data or facts when measurements are per-
formed, and a consistent (average) value can be assigned with
a given error to its physical properties (the distance in the
previous example). We then define objective data or facts as
follows:
Objective data or facts

Those obtained by consistent measurements on the same phe-
nomenon, irrespective of the observer.
How many times should a measurement be performed?

That depends on the type of measurement one performs
and the desired level of confidence (error) we are willing to
tolerate. It may be impractical to walk one kilometer with
well-defined steps many times over, but if we are happy
with an error of, say, 10 meters we could probably get away
with just three strolls.1
1The “error analysis” I have presented here is very simplistic. It is

not intended to capture all the subtleties of the measurement process.
A more in-depth discussion of measurement and error analysis can be
found in, e.g., the book by J. R. Taylor, An Introduction to Error Analysis:
The Study of Uncertainties in Physical Measurements (University Science Books,
1996).
The data are then declared objective because another

observer could repeat the experiment, maybe with yet
another type of instrument, and confirm the previously
recorded results, possibly with a smaller uncertainty.
If this observer finds inconsistent data with respect to
what has been done previously, then the process needs to
restart: either these new data are somehow wrong or they
are affected by some other effect not taken into account or
nonexistent in the original measurement.
Alternatively, the new data may have been able to
uncover a new aspect of the phenomenon not previously
observed because the previous instruments could not reach
that level of accuracy in studying that phenomenon.
Incidentally, this is how many new discoveries in Science
are made: by building new instruments that can reach
farther than the previous ones. A case in point is the tele-
scope built by Galileo that allowed him to see, for instance,
features of the Moon and Saturn’s rings previously unseen
by naked eyes.
As an aside, the above procedure is true also for Quantum
Mechanics and, indeed, by keeping in mind the experimen-
tal way in which quantum phenomena are probed may
dispel much of the “myths” surrounding them.
For instance, when we say that a quantum mechanical
entity is “both a particle and a wave” (as if we could contradict
the “law of identity”) we are actually misinterpreting the
experimental results quite grossly.
To be more specific, take the famous two-slit experiment
in which a beam of particles, say electrons, impinges on
a screen with two narrow openings. On the other side
of the screen there is yet another one (call it screen 2)
that collects the particles that pass through the slits and
measures their position. Every time a single electron hits
screen 2, one observes a dot where the collision occurs,
indicating that the electron behaves as a single particle at
that particular position.
However, once many electrons have hit screen 2, by look-
ing at it somewhat from afar, we see an interference pattern
typical of waves. It is not the single electron that shows wave
properties but all of them together!
In fact, in order to see the wave properties we are forced
to measure many electrons colliding on screen 2, which is
precisely what the scientific method forces us to do anyway:
a single measurement on a single electron would not show the
“wave” behavior!
Note that, at this point in time, we do not know if there
is another cause responsible for the observation of such
an interference pattern of atomic particles (the famous
“non-locality” of Quantum Mechanics): we simply observe
it and describe it quite well with the Quantum Mechanics
equations.
In other words, we have not been able yet to “regress”
one step back in the “law of cause and effect” and probe
some other phenomenon (some other particles or fields
that pervade the Universe?) that could be the cause for
this effect.
Will we ever find such a cause? The scientific successes
of the past give us much hope for the future, but the real
answer is: nobody knows!
Irrespective, this chapter should have conveyed an
important point worth stressing once more. The infor-
mation we collect of the natural world (the objective

data or facts obtained by consistent measurements and
independently of the observer) is limited by how “far” our
instruments can “see.”
By improving our experimental probes or inventing new
ones, we can see farther, towards either the tiniest scales of
the atomic world or the farthest corners of the Universe.
This allows us to discover new phenomena and to make
more sense of others that we have already observed (such
as, possibly, the interference pattern of atomic particles).
Since our interest and our measurements always pertain
to just a particular (extremely small) subset of the whole
natural world (say, the electrons impinging on a screen,
but not their interaction with the surrounding light), that
particular subset of material reality is very often difficult to
isolate completely from the rest.
As such, the information we extract from our measure-
ments always carries a level of uncertainty.
Or to put it differently, there is no such a thing as “abso-
lute certainty” in the knowledge of the material world that
we acquire experimentally. This is true for “classical” or
“quantum” phenomena. It is simply an intrinsic feature
of the measurement process, and applies to all Natural
Sciences with no exception.

• Our instruments can only access a limited subset of the
material world at once.
• Our measurements always carry errors.
• Data about a given phenomenon are objective when

obtained by consistent measurements on the same phe-
nomenon, irrespective of the observer.
• Consistent means that the data from different observers
agree within their respective measurement errors.
• Inconsistent data among researchers require either a re-
evaluation of those data or additional measurements,
so that, ultimately, all data are reproducible within
measurement errors.
5
Observation and Experimentation
Let me linger a bit longer on the concept of repeating an

experiment for the benefit of obtaining objective data.
As I have already mentioned, in order to explore the
material reality around us we need to interact with it either
with our senses or with instruments. The object of study
is then never truly independent of the way in which we
observe it. With this, I mean not just the instrument used to
probe the phenomenon, but also the environment in which
I perform the measurement.
For instance, if I want to test the classical law of dynamics
which says that the force exerted on an object is propor-
tional to its acceleration (variation of velocity in time),

DOI: 10.1093/oso/9780198825623.001.0001
I may track the trajectory of a ball falling under the grav-

itational pull of the Earth. In doing so, however, I cannot
eliminate completely the air surrounding it. The air exerts
a drag on the ball that is proportional to its velocity, not
its acceleration. This effect is even more evident if you let a
feather drop: it will swing slowly toward the ground.
In fact, if I got a ball with the same mass as the feather
and would let them both fall at the same time and from the
same height with respect to the ground, the second law of
dynamics would predict that the ball and the feather reach
the ground simultaneously. This is in obvious contradiction
with the observation that the ball reaches the ground while
the feather is still rocking in midair!
If I make the statement (hypothesis) that the air is
responsible for the disagreement, I can come up with an
experiment in which I place the ball and the feather in
a sealed chamber where I can take out as much air as
possible with a vacuum pump, and then repeat the above
experiment and check the results. If the air vacuum inside
the chamber is really good, meaning very little air remains
in the chamber, the experiment will show that the ball and
the feather indeed reach the ground essentially at the same
time (within measurement errors).
As you can easily see, this type of operation constitutes
a “controlled” laboratory experiment where the environment
(the air) with which my objects of study (the ball and the
feather with the gravitational pull of the Earth) interact is
well under control. My measurement can take into account
not just the error of my instrument (say my eyes). Also the
effect of the environment can be accounted for.
Observation and Experimentation 35
However, such controlled laboratory experiments can-

not always be performed. For instance, in Cosmology, it
is quite difficult to test directly an individual hypothesis
or eliminate the role of the environment, or sometimes
even repeat the experiment on the same object under the
same conditions. The reason is simply because the object
under study does move with respect to Earth, and it will
not return to the exact initial condition of the experiment
in any foreseeable future.
Nevertheless, also these “one-time” observations are facts
that contribute to constructing an economical, logical syn-
thesis that attempts to make sense of apparently diverse
and distinct effects. Such interpretation is what we call a
“theory.”
I will discuss later the role of a theory. But first, let us
discuss in the next chapter an important aspect that makes
the whole Science enterprise work. This aspect relates to
how we, scientists, use the work of our fellow scientists.

• We can never completely isolate the particular phe-
nomenon we are interested in from the influence of its
material environment.
• Controlled laboratory experiments on a particular
phenomenon may not always be easy to perform.
• This means that we cannot always completely know
the effect of the environment on the particular phe-
nomenon we are interested in.
6
The Role of Human Faith in Science
We know that there are eight planets in the solar system;

that light travels in vacuum at about 300,000 km/sec irre-
spective of the relative speed of the source and the observer;
that the ionosphere (the upper layer of our atmosphere)
reflects radio waves, enabling us to communicate over large
distances of the globe; that atoms support discrete energy
levels, and so on.
How do we know all this? I personally never did any
experiment to test any such claims. However, presumably,
someone did, some time ago, and others have repeated
them. These statements were then declared objective data
or facts.

DOI: 10.1093/oso/9780198825623.001.0001
The Role of Human Faith in Science 37
We read them in some textbook or journal article, and,

importantly, we trust the sources of such claims. The word
“trust” is the correct one here: we indeed have faith in
what scientists have done in the past (and are currently
doing) and we do not question the data that they have
generated once these have been declared objective, namely
reproduced by other researchers.
This type of human faith (as opposed to supernatural faith
in God) is an important prerequisite to make the whole
thing work: it would simply be unrealistic to repeat all the
experiments ever done since the beginning of humanity on
anything that pertains to the Sciences.
We simply accept that if some honest, accurate, and
reproducible work has been put into performing such
experiments, then they belong to the collective body of
knowledge.
(When I was studying mathematical analysis as an
undergraduate, our professor made a somewhat similar
statement regarding Mathematics: no one can ever re-
demonstrate all theorems. It is enough that for a given
theorem you know at least one person who has demon-
strated it. I think he was joking, but the statement holds
some truth.)
Of course, it could happen that by repeating the
same experiments, with more precise instruments, new
phenomena may be uncovered.1
Once they have been declared objective, as dictated by
the scientific method, they will also enter into our accepted,
1 Or what we thought were objective data turn out not to be the case
because of faulty experiments.
collective knowledge, and faith in their validity is our only

way to move on to other discoveries.
This type of faith, however, need not be the same as
“faith” in the interpretation of the objective data.
In fact, I need to stress here (and I will expand it later
as well) the important difference between the objective
data (facts) themselves and how we interpret those data. The
interpretation of data is not necessarily objective, namely
independent of the observer.
For a given set of data, there could be different ways in
which scientists make sense of them.
Take, for instance, the attraction of a body to Earth. What
we call “attraction” is a fact, very much in plain sight of
anyone on this planet.
However, we usually go a step further and say that this
attraction is due to a “gravitational force.”
Despite being part of the lexicon of everyone, starting
from elementary school, the name “gravitational force” is
simply an interpretation of the fact that bodies are attracted to
Earth (and attract each other anywhere in the Universe).
Why am I calling it an “interpretation?”
Isn’t it a “fact” that the gravitational force is responsible
for the attraction of any massive object to any other one,
irrespective of whether these objects are confined to Earth,
or anywhere in the Universe?
It all depends on what we mean by “fact” here.
If by “fact” we mean what we have just discussed, namely
that objects attract each other, then we all agree that
this statement is correct. For instance, it has been verified
repeatedly, on objects of different material properties,

shapes, size, etc.
Then I can say that instead of calling it “attraction,”
I have simply given it a new name, “gravitational force,” and
we should understand each other on what we mean. All is
good.
However, if by “fact” we mean that such an attraction
is described by a well-defined (mathematical) concept we call
“gravitational force,” then there is room for debate.
In fact, the concept of “gravitational force,” namely a
vector (like an arrow with a direction and amplitude indi-
cating where the attraction points to and how strong it
is), exists only in classical Newtonian Mechanics, and with
appropriate mathematical modifications, also in Einstein’s
Special Relativity.
However, it is just a mathematical description of the attraction
we have just discussed. Therefore, this particular descrip-
tion can change, and indeed it has already changed to some
extent!
For instance, in the theory of General Relativity, the
attraction among objects is not described by a vector we call
“gravitational force.”
Rather, it can be described and interpreted as originating from
the deformation of the spacetime continuum due to the
presence of masses. Two objects attract each other by trav-
eling the shortest path in this curved spacetime continuum
they themselves perturb.
Therefore, this example shows again that there is a funda-
mental and important difference between the facts we observe
(e.g., the attraction between objects) and the description (or

interpretation) of such facts (e.g., using the concept of a “force”
or “bending of spacetime”).2
Although the collection of objective data may some-
times generate some procedural discussions among scien-
tists (e.g., how many measurements have been performed
in a particular experiment, on how many different objects
sharing the same properties, etc.), oftentimes it is the inter-
pretation of such data that creates many heated discussions
and animosities in the scientific community.
Ultimately, however, experiments (should) settle any
dispute. This last important point deserves much more
discussion than a single sentence. For this, I will devote
more space in the coming chapters where I will explain
what a scientific theory is.
Before doing that, though, let us point out another limit
of scientific inquiry.

• We must rely on, and have faith in the data collected
by other scientists, if these data have been found to be
objective.
• This is because there is no way for us to repeat all the
possible experiments that have been performed from
the dawn of humanity.
2 One could argue that even the word “attraction” is descriptive.
Indeed, we need some “basic language” to communicate with each other

and to point to the phenomena we want to study. This “basic communi-
cation language” is, however, not usually the type of “description” that
I mean here or in the rest of the book.
• This type of “trust” in our fellow humans is, however,

not the same as faith in the supernatural, e.g., God.
• The trust we put in the collection of data does not
translate into “faith” in the interpretation of those data.
• The same data can be interpreted differently.
• Hence, interpretation of facts is not necessarily objective,
namely independent of the observer.
7
Approximate and Limited
Description of Natural Phenomena
In view of our inability to study at once the entire material

reality, our experiments necessarily address only a very small
subset of such a reality.
To this constraint, which I may call an “experimental
limitation,” we need to discuss yet another limitation,
which is related to the description of such a probed reality.
I call this extra constraint a “description limitation.”
To explain this point let me again start with an example.
If I observe a ball falling to the ground, I know by expe-
rience that this phenomenon is not limited to just that
particular ball. I have seen many different objects (and
I have experienced it with myself) that if let go of above the
ground, they return to the ground.

DOI: 10.1093/oso/9780198825623.001.0001
Approximate and Limited Description of Phenomena 43
I can then conclude that all these phenomena, which

occur with so many distinct objects, share some common
thread: no matter their shape, size, color, etc., they all
fall if left alone above the ground. It is then natural to
think that I could summarize all these observations in some
“economical” (preferably mathematical) way.
This logical synthesis accomplishes two tasks: it allows me
(1) to describe what is happening to many, quite distinct
material objects, with just one or a few (preferably math-
ematical) statements, and (2) to predict what will happen if
I perform experiments with other objects.
Here, I have stressed twice the words “preferably mathe-
matical.”
This is because a mathematical description is much more
powerful than what simple words can accomplish in terms
of describing phenomena and predicting new ones.
(As we will see in the next chapters, prediction of new
phenomena is an important and inalienable step in the scien-
tific method.)
The mathematical description itself is always approximate,
and limited to a very specific set of phenomena.
Approximate because there is no way to completely isolate
the particular phenomenon we are trying to describe from
the effect of the surrounding “environment” we are not
interested in. Hence, the description of such a phenomenon
is always incomplete.
For instance, when we say that the gravitational force on
objects of equal mass on Earth produces the same acceler-
ation, we make a strong approximation: we tacitly neglect
the effect of friction that one would naturally have if the
experiments were performed in air.
Limited because this description pertains only to a particular

set of phenomena. For instance, although it would be nice,
we do not describe (yet) the full range of gravitational
phenomena with the same mathematical description we
employ, at the moment, for the quantum ones.
Finally, in order to accomplish the two tasks of describing
and predicting, I need to formulate better the steps that I need
to follow, and which make up a scientific theory.

• We are faced with several limitations when we explore
the material world.
• One limitation relates to our inability to probe the
entire material reality at once: our experimental
instruments are limited.
• The (much smaller) part of the whole material reality
we study with our instruments is also not fully isolated
from the rest we are not interested in.
• Hence, the description of phenomena is always
incomplete.
• In addition, our (mathematical) description of phenom-
ena only pertains to a particular (limited) class.
• Typically, different classes of phenomena have quite
different descriptions, which cannot always be reconciled.
8
Hypothesis
So, now that I have observed a phenomenon, say the attrac-

tion of objects to the ground, and I want to study it, what
do I do?
First, I need to come up with a hypothesis or a set of
hypotheses, namely statements, regarding the objective facts
that I observe.
The hypothesis could be as simple as “There is an attrac-
tion between any object and the ground,” or more gen-
eral “An attractive force is exerted on and by all material
objects in the Universe.” Yet other hypotheses may be “All
species originate from a common ancestor,” “The Universe
was confined to a singularity in space and time, long time
ago,” etc.

DOI: 10.1093/oso/9780198825623.001.0001
The hypotheses need not be mathematically precise

or even extremely specific about the objective data. It
is enough that they are descriptive of what could have
generated such data. For instance, “There is an attraction
between any object and the ground” is descriptive of the
fact that objects left above the ground do fall.
In other words, we say that the hypothesis is consistent with
what we observe.
But is the hypothesis true in the objective sense? Namely,
does it share the same level of confidence as the objective
facts?
Note that being “consistent” does not necessarily mean
“objective.”1 For instance, I could come up with another
hypothesis or a set of other hypotheses to describe what
I observe.
Since air is everywhere on Earth, I could suggest that it
is air molecules which are responsible for the pulling force
on the objects. Without additional tests, I could even claim
that this new hypothesis is consistent with the observations,
since whenever I leave objects above the ground they fall,
and invariably I can breathe every time I do the experi-
ments, which means that air is always present when the
objects fall.
But then, I could test this hypothesis directly by redoing my
experiments in a vacuum chamber and I would see that air
1 Objective, as intended previously, means “independent of the
observer.” As I have already discussed, in the context of the hypotheses,

there may be a dependence on the observer, in the sense that different
observers may come up with different hypotheses to account for the
same facts.
Hypothesis 47
has nothing to do with the fact that the objects always fall
to the ground.
The reader may laugh at the weird hypothesis I came up
with for gravity and indeed, in this particular case, there is
an easy way to check whether it is consistent or inconsistent with
the observations.
But let us now analyze the statement “The Universe was
confined to a singularity in space and time a long time ago.”
Can I test this statement directly? Of course not!
There is no way for anyone to go back in time and observe
even once, let alone many times, this event. This hypoth-
esis seems to be consistent with some observations we make
today (like that of background radiation pervading the Uni-
verse) and it has been the working hypothesis of modern
Cosmology.
But is it a fact? Not at all, and despite what many would
want us to believe, it will never be elevated to the rank of fact,
unless we could build a time machine to go back in time
and check it, or devise a controlled laboratory experiment
in which we reproduce a singularity in space and time, and
observe its consequences!
The event itself is beyond our ability to test it directly;
hence, it is not a fact.
What about “All species originate from a common ances-
tor”? This is again the working hypothesis of modern Biol-
ogy. But can we test it directly and declare it a fact?
Following the scientific method, in order to do so I
would again need to go back in time to when it occurred
(namely, when our first ancestor somehow initiated its
transformation into something else) and test it by direct
observation. Alternatively, I need to devise a controlled
laboratory experiment in which I do observe such a

transformation, say from one species to the one that follows
next in the evolutionary path.
Note that, despite what we hear often, it follows from the
scientific methodology that adaptation (sometimes called
“microevolution”) within a single species is not enough to
prove that “All species originate from a common ancestor”
is a fact. That hypothesis may be consistent with the observa-
tion of adaptation within a single species, but the latter does
not prove directly the former.
So the answer is again, no, the above working hypothesis
is not a fact!
It is the working hypothesis of Biology that attempts to
create an economical, unified way of describing what we
observe now, but by no means is it an objective fact in the
scientific sense.
And I could go on and on with all the other hypotheses
that we come up with in the Sciences. They may be con-
sistent with the objective data they are meant to describe.
However,
Hypotheses are not objective data or facts!
Even the hypothesis considered to be the most obvious

of all, namely “An attractive gravitational force is shared by all
material objects in the Universe,” is simply an hypothesis
that has allowed us to describe (and, mind, really well!) an
enormous amount of data, thus making it perhaps the most
consistent of all the hypotheses I have mentioned. But the
hypothesis itself is not a fact.
Hypothesis 49
As I have already mentioned, in Einstein’s theory of Gen-

eral Relativity there is no concept of a “gravitational force”:
material objects “interact” with each other through the
mutual deformation of spacetime. I put the word “inter-
act” in quotes, because in General Relativity one interprets
the gravitational interaction between objects as the objects
traveling the shortest path in a curved spacetime.
In other words, the word “force” or “interaction” is sim-
ply descriptive of the observations we make. The instruments
we use to make the measurements detect something that
we interpret as (or call) a “force” or “interaction” or “bending
of spacetime.”
Such an interpretation could, however, be modified in
the future as new phenomena are uncovered. It may well
turn out that new data will require extra hypotheses to
describe in a consistent way both these new data and the
old ones.
For instance, when, and if, we are ever able to describe
gravity within the same mathematical structure as Quan-
tum Mechanics , we may very well abandon the hypothesis
that there is an attractive force (or bending of spacetime)
shared by all objects in the Universe, and replace it with
another hypothesis that also takes into account the prob-
abilistic interpretation of atomic phenomena.
Another example is the recent discovery of gravitational
waves that has made big news even outside scientific circles.
In the presentation of this discovery, these waves are
believed to originate from the collision of two black holes.
The collision of these black holes has then supposedly
deformed the spacetime continuum around them, thus
creating the gravitational waves that have been observed in

some laboratories on Earth.
Which part of what I just wrote is a fact, and which part
is simply a hypothesis consistent with such a fact?
Definitely, the detectors used in the experiments have
revealed signals (e.g., the interference pattern of light when
a gravitational wave passes by) with an amplitude greater
than what can be accounted for by other types of effects
(e.g., background noise). What the instruments have mea-
sured is a fact, so waves have been observed.
The data have been collected, experiments have been
repeated, and the results, compared among different teams,
have finally been declared objective facts. We have faith that
the teams of scientists doing this work have all followed the
correct procedure as dictated by the scientific method.
What about the origin of these waves? Are they really
coming from the collision of two black holes some billions
of light years away from us? Is this a fact?
No, this is not a fact!
Despite the nice illustrations of two “black holes” close to
each other and “waves” coming out of their collision that
we see in magazines or newsfeeds, we have to admit that
those are just fruit of the illustrator’s imagination.
Black holes cannot even be observed directly. One can
infer indirectly their possible presence from their effect
on massive objects nearby. So, how in the world does the
illustrator even know what they look like?
The back-hole collision is a hypothesis that through
mathematical calculations predicts a phenomenon consistent
with what has been observed by those instruments on
Earth.
Hypothesis 51
It is the leading hypothesis now widely accepted by the

scientific community, but by no means is it a fact.
Indeed, it could very well be that another hypothesis
that is consistent with the observations, and yet substantially
different from the one of colliding black holes, will turn up
(or is already out there but not considered “mainstream”).
Mind that if this happened, it would not lessen by a bit the
discovery of gravitational waves! It would simply attribute
the description of their origin to some other possible cause.
From a completely different scientific field, consider the
nice pictures of dinosaurs, with their flesh and skin texture
beautifully colored, that we see in textbooks or articles.
Are the flesh and skin we see, or even the movements and
their vocal sounds “reproduced” in many movies, objective
data or facts?
Of course not. Those are again nice illustrations that
attempt to put on paper (or on screen) “educated guesses”
starting from nothing more than fossils.
The fossils themselves are objective data or facts. They are
on plain display for all of us willing to visit a museum.
The rest is an “add-on” that may appeal to our eyes and
imagination, but nothing more than that.
In fact, in order to prove that a given dinosaur has that
particular skin or moves in that particular way, or roars as
in the Jurassic Park movies, we would need to go back in
time and observe all these things directly!
Extrapolating fossils into full-fledged (moving and
breathing) bodies puts us beyond the scientific method and
confines us to the realm of pure speculation.
Finally, let me say that although consistency of a hypothesis
(or a set of hypotheses) with the experiments they are
trying to describe is a necessary condition, it is not sufficient for

it to qualify as a valid hypothesis.
To explain this important point, we need to move on to
what constitutes a scientific theory.

• Hypotheses are statements (not necessarily mathemati-
cal) we make to describe the observed material world.
• Hypotheses per se are not objective data or facts.
• Hypotheses may be consistent or inconsistent with the
observations they are meant to describe.
• It is not always easy to check directly, by performing
an experiment, whether a hypothesis is (in-)consistent
with observations.
• In particular, when hypotheses refer to past events they
are definitely not checkable directly, unless we are able to
reproduce in the lab the phenomena they are meant to
describe.
9
Theory
Once we have settled on one or more hypotheses, our task

as scientists has just begun. As I discussed previously, we
can come up with many hypotheses for the same body
of observations. Some of them are easy to eliminate by
direct measurements, and hence deemed inconsistent with
the data, but we could end up with a few competing ones
that superficially all look consistent. How do we distinguish
among these?
We need to come up with a comprehensive set of descrip-
tive statements that allow us to (1) describe what we have
already measured, and (2) make predictions of new phenomena

DOI: 10.1093/oso/9780198825623.001.0001
yet to be observed. Clearly, these statements need to build

logically on the hypotheses we have made.
The set of descriptive statements that accomplishes those
two tasks, together with the hypotheses they build upon, is
what we call a theory.1
Note first that, like the hypotheses, the descriptive state-
ments of a theory need not be mathematical. Let us again
consider gravity as an example. My theory of gravity could
simply be formulated in these two statements: “There is an
attraction between any object and the ground” (hypothesis)
and “Every time I leave an object above the ground to its
own devices, it will fall toward the ground” (prediction).
These two statements alone are enough to accomplish
what I was after: (1) to describe in an economical way
(namely, with a small amount of hypotheses) all the
observations I have made in my lifetime regarding objects
left above the ground, and (2) to predict that if another
observer leaves an object, which I have never thought of,
above the ground to its own devices, that object will also
fall.
This last step is really crucial: with this one prediction,
which can be tested experimentally by independent observers, I can get
rid of other hypotheses that could have been put forward
(say, the air one I came up with before).
However, suppose now I want to know more precisely
the time it takes a given object to reach the ground once it
1 In certain literature, e.g., in Physics, the word “theory” is some-
times replaced by the word “model.” The meaning attributed to both

words and to their scientific role is the same, so I will not distinguish
between them.
Theory 55
has been released from a given height. This requires further

hypotheses, namely that if we could eliminate any possible
interference from the environment (say, the air or other
obstacles), the “attraction” we guessed could be described
by a vector that we call “force” that pulls the object toward
the ground (direction of the vector) with a strength pro-
portional to the acceleration of the object (magnitude of
the vector).
This statement, usually called the second law of dynam-
ics, is a mathematical description that allows us to make incred-
ibly precise predictions on the motion of objects subject to
gravity, and not just on Earth. For instance, when applied to
planetary objects (and written so that the magnitude of the
force between two objects is proportional to the product
of their masses, and inversely proportional to the square
of their distance) it made it possible to describe the orbits
of planets in the Solar System quite accurately, and even
the remarkable prediction of the existence of the planet
Neptune!
The reader can easily grasp then why Mathematics is
the ideal (and universal) language in which to write the
descriptive statements of a theory: it allows us to perform
accurate, experimental checks of specific predictions that
come out of that theory. Since Mathematics is universal,
such checks can be planned by anyone on this planet willing
to test my predictions.
By limiting the theory to just words (like “objects always
fall to the ground”) we could not have made, e.g., the
prediction of an extra planet in the Solar System, and taken
advantage of the predictive power of a mathematically for-
mulated theory in innumerable other cases as well.
It is also important to remember that theories are built

upon a finite set of hypotheses, which constitutes their foun-
dations.
The finiteness of this set of hypotheses defines the reach of
a theory in terms of the phenomena that it can describe and
those it can predict.
Two theories built upon different sets of hypotheses are
fundamentally different. I will expand on this point in a sub-
sequent chapter.
For now, I want to make it clear that the fact that a theory
(or model) is built on a limited set of hypotheses is not just
an abstract concept with only “academic” relevance. It has
practical consequences as well.
Consider for instance, the hypotheses that go into
Quantum Theory (and if you want to extend it to rel-
ativistic phenomena, you may consider Quantum Field
Theory).
Since this is not a technical textbook, I will not report
these hypotheses here. In fact, the only thing we need to
know about them is that Quantum (Field) Theory was
conceived originally to describe phenomena pertaining to
(sub)atomic particles.
However, you could argue that since subatomic particles
constitute the building blocks of all matter, we should be
able to describe all phenomena with Quantum Theory,
including those that pertain to, say, the macroscopic organ-
isms of Biology.
For instance, I should be able to use Quantum Mechanics
to describe how two or more organisms interact with each
other in their environment.
Theory 57
This “reductionist” approach is often advocated by some

physicists. However, while it may have some merit as a
“matter of principle,” it misses completely the practical role
of a theory, in this case Quantum Mechanics.
I will stress more this point later in the book using differ-
ent examples. For now, in order to show how this practical
limitation occurs, let us follow this reasoning.
Can we write down the quantum mechanical (Schröd-
inger’s) equation describing all the subatomic particles
constituting a virus? In principle, yes (if we have enough
information about the types of atoms making up the
structure of a virus).
Can we then write down the same equation for a virus
attacking a cell and injecting its RNA into it? Yes, of course.
This equation can, in principle, be written.
Is this quantum mechanical theory of virus–cell interac-
tion of any help? Absolutely not!
There is no way for us to even remotely come close to solv-
ing this equation with such a gigantic number of particles,
even with the aid of the fastest supercomputers on Earth.
So, while in principle we may formulate the interaction of
a virus with a cell using Quantum Mechanics, this exercise
is of no practical value.
Therefore, in addition to the fundamental limits imposed on
Quantum Mechanics by its own (limited) set of hypothe-
ses, we have also found a practical limitation of Quantum
Mechanics in tackling some phenomena that, in principle,
should be within its reach.
Note that I am not saying that no phenomena in
Biology fall within the practical description of Quantum
Theory. In fact, some specific effects may even require

Quantum Mechanics to make a quantitatively predictive
description possible, e.g., when chemical reactions occur in
some biological complexes due to the interaction of some
molecules with light or heat.
However, in all these cases, we isolate those (extremely)
small parts or molecules (compared to the whole organ-
ism) where these reactions occur and study, quantum
mechanically, what happens to these small parts
(molecules).
We are definitely not considering the biological organism
in its entirety!
This limitation, however, has no bearing on the appli-
cability of the scientific method, so long as the quantum
mechanical description of such a small part of the organism
leads to testable predictions.
This last point is decisive: we are free to choose whichever
subset of the whole system we want to describe (in the
present example, some of the organism’s molecules), and
free to choose whatever theoretical approach we want.
However, to have any meaning in terms of information
we can extract from the material world, this description
must lead to testable predictions, or it will always be con-
fined (at best) to the realm of hypothesis.
The practical limitation I just discussed is not confined
only to Quantum Mechanics, but it is intrinsic to any the-
ory (or model) we come up with to describe and predict
phenomena.
For instance, referring once more to the Biology example
of describing a virus attacking a cell, it would still be imprac-
tical to describe all the particles of the virus, the cell, and
Theory 59
their liquid environment even if we used classical Newto-

nian Mechanics. There is no way we could solve Newton’s
equations of motion for all of the virus’s interaction with a
cell over the time scales when the virus injects its RNA into
the cell nucleus.2
We simply do not have the computational resources to
do so.
And even if we did have those resources, the question we
should ultimately answer is: What information do we want to
extract from our theory that makes us learn more about
this particular phenomenon?
In other words, which predictions can we obtain from such
an atomic-scale description that cannot be obtained from
a simpler theory (model) that simplifies the problem, e.g.,
by treating the virus and the cell as a system with few
parameters and few equations of motion?
As I said, ultimately the choice of what theory (or model)
to use to describe a particular phenomenon is left to the
researcher, provided the essential role of making predic-
tions out of that theory is not forgotten.
All this should also further clarify the distinction
between the phenomena that we observe and their theoretical
description.
First, we observe some phenomenon, say, the falling of
objects to the ground when left to their own devices. And
this is the phenomenon we want to describe.
2 Note that this method, typically called “classical molecular
dynamics,” is successfully employed in a variety of fields, such as Mate-

rials Science and Chemistry, when the number of particles is small
enough to allow for an efficient computation of the equations of motion.
How we describe it is a completely different matter!

I can describe it in simple words, with simple Mathemat-
ics from Newtonian Mechanics, or I can use the non-linear
equations of General Relativity, or . . .
The list of possible theoretical descriptions goes on (and
will continue to grow with time). The phenomenon is what
it is, while its description may change over time.
By changing the description, however, we introduce new
sets of predictions, and if these predictions are tested exper-
imentally, new phenomena may emerge (e.g., bending of
light from massive objects as predicted by General Rel-
ativity, but completely outside the range of Newtonian
Mechanics).
The progress of Science is then a continuous succession of
observations, theoretical description of these observations,
predictions of new phenomena, and again observations of
these new phenomena, if the predictions turn out to be
verified by experiments.
We start with experiments and we end with experiments.

• Scientific theories are a set of descriptive statements (not
necessarily mathematical) that describe phenomena
already observed, and make predictions of new phenomena
yet to be observed.
• Theories that employ mathematical statements have a
greater predictive power than non-mathematical the-
ories.
• Theories need to build logically on their foundational
hypotheses.
Theory 61
• The number of hypotheses is finite, so that it defines

the reach of a theory regarding the phenomena it can
describe and those it can predict.
• This is true also in the practical application of a theory.
Even if a theory could, in principle, be applied to a num-
ber of different phenomena, unless it allows us to make
predictions, it is practically useless.
• The reason is that, without predictions, we are left with
only a bunch of descriptive statements, whose validity
cannot be tested experimentally.
• Experiments are (should be) then both the beginning and
the end of any scientific enterprise.
10
Competing Theories
I have discussed what a scientific theory is. I have stressed

that it is not enough that it describes the phenomena it was
designed to tackle. It also needs to predict new phenomena.
Predictions elevate a set of hypotheses from pure specu-
lation to the testable construct we call theory.
Predictions also offer the necessary (and testable) checks
to distinguish among competing theories.
Suppose I come up with two distinct theories for a given
set of observations. Suppose also that those two theories are
built on quite different sets of hypotheses.
Which one should we choose?
If the two theories describe the same phenomena and lead
to the same predictions, then there is no specific criterion
to decide between the two: both theories are scientifically
acceptable.

DOI: 10.1093/oso/9780198825623.001.0001
Competing Theories 63
Faced with this situation, which one should we choose over

the other?
I highlighted the word “should” because, in this case, we
need to make a couple of extra considerations.
First, if we are faced with two distinct choices that lead to
the same outcome, we tend to follow the “easiest” one. In
the case of a scientific theory, this means choosing the one
built on the least number of hypotheses.
This criterion, sometimes called “Occam’s razor,” has
another obvious advantage: the more hypotheses we intro-
duce to build a theory, the more likely it is that, with time,
new experimental data will render one of those hypotheses
inconsistent with the data.
The second consideration is less edifying for the whole
scientific community. Since Science is made by humans, it is
not immune from their faults and shortcomings, especially
pride and envy.
In recent modern times, these shortcomings have been
compounded by the substantial amount of money Science
has required to progress, but pride and envy have always
been responsible for the major animosities among scientists
of any period.1
The choice then between competing theories has been,
unfortunately, sometimes “settled” by means that have
little to do with the logical or mathematical simplicity of
1 See, e.g., the controversy between Newton and Hooke on the

theory of colors, or the one between Pasteur and Pouchet on the
spontaneous generation of life. The interested reader could learn more
about these cases from the book edited by P. Machamer and M. Pera,
Scientific Controversies: Philosophical and Historical Perspectives (Oxford University
Press, 2000).
a theory, and more to do with the “clout” or support a

group of individuals receives from the community. When
this happens, Science does not necessarily gain.
The situation is definitely easier to sort when the two
theories produce at least one or more different predictions. In
this case, experimental evidence should settle the question,
and declare “victorious” the theory that correctly predicted
the outcome of the experiment.
This example shows once more the central role of exper-
iments or observations and makes it clear that (in an ideal
world):
Science is not a democracy: Nature rules!
It doesn’t matter if almost all the scientists in the world

agree on a theory, while an “unknown” chap is the only
one with a competing theory. As much as pride can be
an obstacle, that majority of scientists must give up their
theory (!) if experiments do not confirm it.

• If two scientific theories are based on different sets of
hypotheses, but make the same predictions, there is no
obvious criterion to distinguish between them.
• However, if that is the case, we tend to favor the theory
based on the smallest number of hypotheses (Occam’s
razor).
• This is because the smaller the number of hypotheses,
the less likely it is that, with time, new experimental
Competing Theories 65
data will render some of those hypotheses inconsistent

with the data.
• Although Occam’s razor favors the simplest among
two competing theories, the choice between the two
can often be dictated by human factors that have noth-
ing to do with the simplicity of a theory.
• If the theories make different predictions, then experi-
ments should settle the issue.
11
Can One Theory Be “Derived”
from Another?
I want to make another important point, which superfi-

cially may seem just an exercise in scrupulosity, but which
should better clarify what a theory is.
We often hear, and we are even taught in some Physics
classes, that Newtonian Mechanics can be “derived” from
Einstein’s theory of Special Relativity as its “limit,” when
we consider speeds much smaller than the speed of light.
From a purely computational/quantitative point of view,
this statement is correct. For instance, if I want to determine
the speed of a train with respect to me and use Newtonian

DOI: 10.1093/oso/9780198825623.001.0001
Can One Theory Be “Derived” from Another? 67
Mechanics to make such calculation, I would obtain prac-

tically the same result if I used the equations of Special
Relativity in their limit of “low speed.”
This is because the speed of light (about 300,000 km/sec)
is many orders of magnitude larger than the typical speed
of a train (even the fastest trains on Earth cannot go too
far beyond speeds of about 400 km/h). Therefore, the ratio
between the typical speed of a train and the speed of light
can be, for any practical purposes, assumed zero.
Can we then say that Newtonian Mechanics is a “limiting
theory” of Einstein’s theory of Special Relativity, in situa-
tions in which the speeds of the material objects involved
are much smaller than the speed of light? It depends on
what we mean by “limit of a theory.”
Following the above example, we need to recall that
there is one important hypothesis that we make in Special
Relativity that is not part of Newtonian Mechanics: the speed
of light is the maximum speed achievable by any material
object, and is a constant irrespective of the speed of the light
source relative to the observer.
Such a hypothesis is key to developing Special Relativity
and has allowed us to abandon the notion of an “ether”
pervading the Universe and serving as the medium through
which light travels.
It is therefore not completely correct to say that one
theory (e.g., Newtonian Mechanics) is a “limiting” case of
(or is “derived” from) another (e.g., the theory of Special
Relativity), because the two have different sets of hypotheses,
and there is no “limiting procedure” that eliminates one or
more hypotheses (e.g., the constant speed of light): either
you have them or you don’t!
This is true for all theories we construct. For instance,

Quantum Mechanics has many additional hypotheses com-
pared to Newtonian Mechanics (or Quantum Electrody-
namics compared to Classical Electromagnetism), so the
latter one is never truly a limit of the former. Similarly,
General Relativity contains a principle (that equates a grav-
itational field to an accelerating reference frame) that is not
part of Special Relativity, and so on.
Again, the calculations one performs in certain limits (e.g.,
speeds much smaller than the speed of light) within one
theory may agree within certain errors with those obtained
from another. However, since they are built on different
hypotheses, the two theories are nonetheless fundamentally
and conceptually different.

• A scientific theory can never be truly considered as the
“limiting case” of another.
• This is because different theories are always built on
different sets of hypotheses.
• No “limiting procedure” can eliminate one or more
hypotheses: either the theory has them or it doesn’t.
• This does not mean that some calculations performed,
and some predictions made, with one theory cannot
agree with those obtained with another.
• They may actually agree quantitatively within certain
limits, even though the two theories are fundamentally
different.
12
Verifying or Falsifying? And What?
Starting from a body of experimental evidence scientists will

move on to construct a theory based on a set of hypotheses
(statements, preferably mathematical). That theory will
make a series of predictions so that it can be . . . what?
We hear a lot about the falsification of a theory. The most
outspoken proponent of such a requirement is the late
philosopher Karl Popper.1
Falsifiability means that to be declared scientific a theory
needs to produce one or more predictions that can be put
1 Karl Popper, The Logic of Scientific Discovery (Hutchinson Education,

1959).

DOI: 10.1093/oso/9780198825623.001.0001
to experimental test. If one of these predictions turns out

not to be in conformity with the experimental results, then
the theory is falsified.
But what does it mean that the experimental results are
not in agreement with the prediction(s) of the theory?
Let us go back to the example of discovering the planet
Neptune that was predicted by Newtonian Mechanics.
Does this mean that Newtonian Mechanics has been
verified by that observation?
Not quite. We can conclude from that particular exper-
iment that the hypotheses of the theory, and its set of
predictive statements, are consistent with the observations to
a level that allows us to describe not just phenomena on
Earth, but also this new discovery at the planetary level.
In other words, we have expanded the descriptive power of the
theory to a larger set of phenomena. Any other predic-
tion that is successfully tested experimentally does precisely
this: it expands the descriptive range of the theory.
Although we may loosely use this term, a theory per se
cannot be truly “verified,” because no amount of predictions
that turn out to be corroborated experimentally can give this
theory the status of an objective “fact.”
Therefore, while the predictions themselves may be veri-
fied (corroborated) experimentally, the whole theory is defi-
nitely not.
In other words:
Theories are not objective data or facts!
Indeed, from the previous discussion we can summarize the

role of a theory simply as follows:
Verifying or Falsifying? And What? 71
Theories
(i) describe a limited set of phenomena, and
(ii) predict new phenomena.
Again, the description may not be mathematical, but a

mathematical description affords a predictive power that
words alone cannot accomplish.
Theories are designed to describe only specific phenomena
of specific natural events. This is the essence of Plato’s “to
save the phenomena” I mentioned in this book’s Preface:
it is enough for the theory to be consistent with the specific
phenomena it is meant to describe, and be able to predict new
phenomena.
That’s all. Nothing else is required from a theory.2
When we say that a theory is “universal,” what we
mean is that it is developed to account for most (if not
all) phenomena that could be clustered in a particular class.
For instance, Newtonian Mechanics has been designed to
describe all phenomena related to the motion of classical
objects. It was not conceived to describe, say, quantum
phenomena.
Even if we could, as nowadays it is hoped for, come
up with a single theory that accounts for the (currently
known) three forces of Nature (gravity, electroweak, and
strong force), that theory would not be a “theory of every-
thing”!
2 See Pierre Duhem’s Essays in the History and Philosophy of Science (Hackett
Pub. Co., 1996) for further discussions on this important point.
It would simply be a theory that describes a particular

(albeit vast) set of phenomena pertaining to those interac-
tions. It would certainly leave open loads of other questions.
For instance, where does the non-locality of Quantum
Mechanics originate from? Are space and time epiphenom-
ena of some other phenomenon? Better yet, other ques-
tions and new phenomena will more than likely emerge
when new instruments that greatly expand the reach of our
senses are devised.
Suppose now that the planet Neptune was discovered
but not exactly where the theory predicted it, say some
distance away.
Now, should we say the prediction is inaccurate and
hence the theory is false?
Not necessarily. We could conclude, for instance, that
some other effects have contributed to the discrepancy.
Maybe imperfect measurements performed from Earth are
the culprit, or the calculations could not be done with
very high accuracy due to numerical errors, or maybe
some extra effects, not yet accounted for by our theory,
need to be considered to have a better agreement with the
experiments.
Given the body of other successful predictions, we would
certainly not throw away Newtonian Mechanics for this.
After all, we did find Neptune!
But again we would not say that we have “verified” such
a theory. We would only say that within a certain error
(whose origin we may not fully understand yet) the data
are consistent with the predictions.
Finally, let us now imagine the worst-case scenario: we
did not find Neptune, neither where the theory predicted
Verifying or Falsifying? And What? 73
nor anywhere such a planet could be seen within our exper-

imental capabilities.
The prediction did not materialize. Something in the
theory is inconsistent with the data at a very “critical” level.
What exactly should we do about it? Is the theory itself false?
Of course not!
It only means that we have uncovered a limit of the
theory, namely that it cannot describe certain phenomena that
we originally thought it could.
To render this point clearer, let us consider Einstein’s
incredible prediction (at that time) based on the theory of
General Relativity that a ray of light bends near a mas-
sive object. This phenomenon was later observed, although
Newtonian Mechanics couldn’t account for it.
This outcome did not result in scientists tossing out
Newtonian Mechanics. In fact, we still apply it very success-
fully when effects like the one above are negligible. We just
uncovered one of its limits.
Similarly, it was found that the speed of light was the
limiting speed of all matter. Newtonian Mechanics did not
contemplate such a limit.
In fact, absence of such a limit in Newtonian Mechanics
may lead to paradoxes due to the superluminal (exceeding
the speed of light) transfer of information between objects.
Such paradoxes can be resolved within the theory of Spe-
cial Relativity or General Relativity, and simply show that
Newtonian Mechanics cannot be applied in those contexts.
However, when phenomena involve speeds of objects
that are much smaller than the speed of light, we safely
employ Newtonian Mechanics to describe such phenom-
ena and make predictions.
Ultimately, it is the experimental test of such predictions

that confirms for us whether the choice of Newtonian
Mechanics was reasonable to begin with, or whether we
have uncovered new effects beyond its reach.
In summary, we do not require the predictions of a the-
ory to “verify” or “falsify” the theory itself, but rather to
determine its limits or to distinguish between competing
theories.

• Theories are not objective data or facts.
• They are meant only to describe phenomena and predict
new ones.
• If among various predictions of a theory one of them
does not materialize (it is not verified experimentally),
then that theory is not “falsified.”
• This outcome only means that we have uncovered a
limit to such a theory; namely, this theory cannot be
applied to some phenomena we previously thought it
could.
• Even though a single theoretical prediction can be verified,
namely corroborated experimentally, a theory per se can
never be “verified,” because no amount of predictions
that turn out to be corroborated experimentally can
give that theory the status of an objective “fact.”
13
Don’t Be a Masochist!
Following up on the discussion of the preceding chapter,

it is worth noting that although we could use Einstein’s
theory of General Relativity to describe phenomena that
are also within reach of Newtonian Mechanics, this would
come at an incredible cost in conceptual effort and compu-
tational cumbersomeness.
Newton’s equations are much easier to handle than Ein-
stein’s equations, and the theoretical framework of New-
tonian Mechanics is more intuitive than that of General
Relativity when we are faced with phenomena that we
experience in our everyday life.
Here, you see again Occam’s razor at work: for the set
of phenomena where both Newtonian Mechanics and Ein-
stein’s General Relativity are equally valid, we choose the

DOI: 10.1093/oso/9780198825623.001.0001
theory that has the least number of hypotheses and pro-

vides the simplest description.
To see this “principle” of simplicity at work in an even
more compelling way, let us first analyze an example that
is frequently misunderstood, not just by lay people, but also
within the scientific community.
We often hear the following statements: it is “obvious”
that the Ptolemaic system is “wrong” in placing the Earth in
a static position with the other planets and the Sun moving
around it. Instead, the Copernican system is the “correct”
one, in which the Sun is fixed, with all the other planets,
including the Earth, moving around it.
It may come as a surprise to some, but both systems are
wrong in some sense, and correct in another.
So, why the apparent contradiction? And, why is the
Copernican system the preferred one?
To answer these questions let’s start by saying that I, and
most of you I am sure, have personally never seen the Earth
moving around the Sun (if we do not count science fiction
movies or documentaries on the subject).
In fact, most of the time, we observe the motion of the
Sun, or any other planet, from the Earth, and in doing so,
of course, we are using the latter as the reference frame.
On the other hand, when we imagine sitting on the Sun
(“imagine,” because we cannot actually do this experiment
of relocating to the Sun without dying!), then we use the
Sun as the reference frame, and compute the orbits of all
the planets, including Earth, from that reference frame.
We can then understand in what sense both the Ptole-
maic and the Copernican systems are wrong.
Don’t Be a Masochist! 77
If by Ptolemaic system we mean that we choose the

Earth as an absolutely still reference frame (namely, one that
is motionless with respect to everything else in the Uni-
verse!), and by Copernican system we mean that the Sun is
instead an absolutely still reference frame, then we make the
same fundamental mistake.
In our physical description, motion of an object has a
meaning only after we have decided which point of reference
to take. In other words, we first choose the reference frame,
and only then can we say that an object moves at a given
speed, and follows a given trajectory with respect to such a
reference frame.
It does not make any sense for us to say that any par-
ticular point in the Solar System, or in the entire Universe
for that matter, is “absolutely motionless,” because the next
question we should ask is: with respect to what?1
Once this point is settled, we can now understand the
(practical) reason for the choice of the Copernican system
over the Ptolemaic one.
The Copernican system offers an incredible advantage
compared to the Ptolemaic one, if we are interested in
describing the motion of the planets in our Solar System.
The orbits of the planets as viewed from a reference frame
whose origin is the Sun are simpler (ellipses) than those we
would obtain by describing their motion from the Earth.2
1 Also the images taken from spacecrafts or satellites show the
motion of the Earth, the Sun, and all the other planets relative to (mea-
sured from) those spacecrafts or satellites.
2 To be precise, the Sun is not at the center of the ellipse, but it would
be placed at one of its focus points.

That’s all. There is nothing more to it! One reference

frame allows us, literally, to do calculations on the motion
of the planets in a much simpler way than the other.3
Now, suppose you want to describe the motion of all the
other stars in our galaxy. If you choose the Sun as the origin
of the reference frame, the orbits of those stars would be
much more complicated than if you chose the center of the
galaxy as the reference. And so on.4
To exemplify further the fact that we tend to choose the
simplest description out of many, equally valid ones, let
us use a completely different example. Let us consider a
problem from Biology: the motion of a colony of ants as
they find food.
This example will also allow me to reconsider the
“description limitation” I anticipated when discussing the
approximate and limited nature of our explanation of
natural phenomena.
You can actually check this example at home if you are
willing to have an ant infestation, or, like me, you may have
observed it as a kid while playing with these insects.
3 In fact, all of the pictures of the entire Solar System, with all
its planets neatly tracing their elliptic orbits around the Sun often
found in textbooks, documentaries, etc., are representations of what a
hypothetical observer outside the Solar System would see were he at rest
with the Sun.
4 Also the stellar parallax observed from Earth (namely, the motion
of the stars with respect to Earth, when their position is measured at

two different days of the year—preferably at six months’ distance) is
not simply a “direct proof” that the Earth moves around the Sun. It
shows the relative motion of the stars, including the Sun, and the Earth,
as measured, in this case, from the reference frame Earth.
Suppose you are interested in studying the way a colony

of ants is able to find food as they come out of their nest.
You will notice that the colony will first send out a few
“explorer” ants to search for food in the environment sur-
rounding the nest.
We then observe that if we give them some time, a famil-
iar black path of ants forms, in which ants go back and forth
from the nest to a food source, e.g., bread crumbs lying on
the floor.
As I mentioned previously, since ants and their environ-
ment, including the materials making up their nest, are all
made of atomic particles, I could use Quantum Mechanics
to study this phenomenon.
I would immediately realize this is a hopeless (some
would even say crazy) proposition.
My next choice would then be to apply classical New-
tonian Mechanics to describe the phenomenon. In this
case, I would need to ignore the quantum structure of the
atoms in all the ants and their environment and treat those
atoms as classical particles following Newton’s equations of
motion.5
I would quickly realize that the number of particles
whose motion I need to consider is so gigantic that no
available supercomputer on Earth would be able to help
me with this.
At this point, I might consider this phenomenon from a
different perspective and advance a simpler model practically
solvable.
5 See Chapter 9, footnote 2.

I would assume (first hypothesis) that the ants use some

form of memory (pheromones) to “communicate” and
navigate their surroundings.6 In this way a path from the
nest to the food source is chosen by the colony.
I then assume (second hypothesis) that the ants’ atomic
structure or that of their environment is irrelevant to
describe the ants’ motion.
After all these approximations, I may therefore model
each of them as “structureless” point entities following a
very simple equation of motion (with memory) as they
travel a network representing the paths they can possibly
choose.
This description is practically easier to handle than using
Quantum Mechanics or Newtonian Mechanics for all the
atoms constituting the ants and their environment. I need
to solve a considerably smaller number of equations.
Most importantly, it allows me to make predictions that
I can test experimentally.
For example, I can ask the question: out of the vast sets
of possible paths, which one is chosen by the ants?
Using my model, I predict that if I simulate the following
situation, the ants choose the shortest path.
Let us perturb the ants’ path already formed from the nest
to the food source by, e.g., placing somewhere along the
path, an obstacle tall enough that now the ants have only
two possible path choices to move around it (one longer than
the other).
6
Memory-dependent motion is indeed a typical feature of physical
systems.
If the ants don’t lose interest in their food, right after

the addition of the above obstacle across their path, they
will explore both options (the short and long paths). Given
enough time, I would observe that the shortest path is
chosen, and a black trail of ants around the obstacle will
clearly show it.
I therefore chose the simplest description possible that
allows me to make a well-defined prediction.
In addition, I can now evaluate the hypotheses that I
made and can conclude that the assumption that ants use a
memory mechanism to communicate is consistent with what
I predicted and observed.
I can also discriminate the hypotheses of my model from
other hypotheses that use different mechanisms.
For instance, another theory could claim that ants some-
how can visually see the food source directly from their nest
with some special optical means. In other words, the ants
have a global awareness of their surroundings, a sort of “ant
binoculars” that allows them to see beyond any obstacle.
In this case, ants would not necessarily choose the shortest
path.
Therefore, if another researcher puts forward a different
description, with very different hypotheses, then their alter-
native theory should at least make the same prediction as
mine to qualify as consistent with the phenomenon observed.
All this to say that even though we have simplified con-
siderably the description of the “ants search problem,” this
simplification has allowed us to describe what we observe and
predict a new aspect of the phenomenon we study.
This simplification, from the (possibly) more “accurate”
(“fundamental”?), but utterly impractical quantum or
classical-particle description of either Quantum Mechanics

or Newtonian Mechanics, has accomplished the goal we
were after: extracting information from the natural world
around us.
In summary, we choose a particular descriptive model
among competing and equally valid ones according to the
simplicity with which it allows us to make predictions and
measure (or make calculations about) the phenomena we are
interested in, not because one is more “correct” than the
other.
You should agree with this point, unless you are a
masochist!

• The practical aspect of a theoretical description cannot
be underestimated, since we typically tend to favor the
choices that “make our life easier.”
• Therefore, the choice we make of a theoretical descrip-
tion, among different and equally valid ones, depends
on how easy it allows us to make predictions and con-
ceive of new measurements.
• There is no such a thing as a more “accurate” or “fun-
damental” theory if it does not allow us to compute
testable quantities.
14
“Consensus” in Science?
What Is That?
We often hear that in this or that scientific community

there is a “consensus ” on a particular interpretation of phe-
nomena or a theory. For instance, there is a clear agreement
in Physics that Quantum Mechanics is the best theory to
describe and predict phenomena at the atomic or sub-
atomic level.
Does this mean that Quantum Mechanics is “true” or
“correct”? Should we even seek to come up with other
theories?
Following the exposition in this book, we can certainly
say that Quantum Mechanics is, at the moment, the best
description we have of a particular class of phenomena.

DOI: 10.1093/oso/9780198825623.001.0001
However, it would not be a stretch to think that in the

future another theory could supersede Quantum Mechan-
ics. Such a theory should be able to describe phenomena
that we have observed so far, but also predict new phenom-
ena that the present formulation of Quantum Mechanics is
unable to describe.
After all, scientific progress has always happened in this
way. For example, a change of scientific “paradigm” (as
Thomas Kuhn would call it1) could come from the inabil-
ity of Quantum Mechanics to predict some unexpected
phenomenon, which, therefore, falls outside its realm of
description. If that were the case, it would require a new
foundation; namely, we would need to add to or change some
of its hypotheses.
And even if that were the case, we would not throw away
Quantum Mechanics all together. We would still use it for
the class of phenomena for which it is valid.
We would simply avoid its application outside of its reach,
as we would not use Newtonian Mechanics to describe, e.g.,
the bending of light from massive objects.
You may be then tempted to think that the scientific
community would welcome such a change of paradigm
with open arms. Far from it!
Despite their reputation as an open-minded bunch, sci-
entists offer a lot of resistance to change. This is both a
blessing and a curse.
It is a blessing, because by resisting a new theory and
putting it to strict tests, it allows the scientific community
1
Thomas Kuhn, The Structure of Scientific Revolutions (University of
Chicago Press, 1962).
“Consensus” in Science? What Is That? 85
to “filter out” ideas or proposals that do not correspond to

objective reality. On the other hand, it is a curse, when it
slows down progress to an almost standstill for a while for
reasons that are not objectively scientific.
This outcome is particularly felt when a small group of
(extremely vocal) scientists promote their own ideas and
create their own “circle” of like-minded individuals.
Alternative ideas or theories are then kept at a distance,
or worse denigrated. This behavior makes it difficult for
newcomers to break into the mainstream of a field of study,
even if they bring valuable new suggestions.
I personally think that the situation has become even
direr in past decades due to the financial pressures put on
scientists, especially the young, untenured ones.
Tenure2 at research institutions is oftentimes based on the
ability of a scientist to receive funding from granting agen-
cies, to publish in “high-impact” journals, to be recognized
by their peers, etc.
This compels many young scientists to remain within the
boundaries of the “establishment” (or follow the latest “hot
fad” of the moment) to avoid making waves that could have
a negative impact on their promotion.
This, of course, does not mean that new ideas cannot get
through, and finally emerge if they are proven to corre-
spond to reality. It simply means that there is an extra layer
of resistance that needs to be shed before that happens.
2 Academic tenure means that the holder cannot be fired for the
ideas they support, even “controversial” ones. Clearly, it does not mean
they cannot be fired if they commit a felony.
This aspect of “scientific guerrilla” is not just of recent

memory.
As an example from the past, consider Boltzmann’s
kinetic theory of gases that assumed the existence of
microscopic objects called atoms and molecules. Boltz-
mann’s contemporaries (especially physicists) were far
from thrilled at such a hypothesis. It took further devel-
opments and, importantly, several experimental tests, to
render the kinetic theory of gases “acceptable” to the wider
Physics community.
Other similar examples were referenced in Chapter 10’s
footnote, and show, as pointed out earlier, that this behav-
ior is part of the fabric of Science. Being a human enterprise,
it is not exempt from human weaknesses!
Ultimately, however, progress happens and even lonely,
isolated voices in the vast cacophony of statements pro-
duced by the community will break through. This happens
precisely because of the existence of an objective reality beyond us,
which provides the ultimate judgment to this enterprise.
These thoughts can be summarized in a single sentence:
Nothing is ever settled in Science!
This sentence holds a very important truth that we

should always bear in mind. “Consensus” among scientists
does not mean that everything is “settled” in a field. This
quality of “certainty” is not proper to Science.
When “scientific consensus” is invoked, it often reflects
some other “forces” at work in those who pronounce it.
“Consensus” in Science? What Is That? 87

• “Consensus” in Science, when it pertains to the interpre-
tation of data, does not exist.
• Although data or facts may stand the test of time, their
interpretation is always susceptible to revisions.
• Indeed, nothing is ever settled in Science.
• If a group of people claim “consensus” on the interpre-
tation of some data, then typically other “forces” are at
play that have nothing to do with Science.
• This “consensus” may have the unwanted outcome of
unwittingly slowing down progress.
• Ultimately, however, scientific progress occurs be-
cause of the existence of an objective material reality
that is the final judge of the validity of a scientific
statement.
15
Flow Chart of the Scientific Method
We have now discussed the complete method that is at the

core of the Natural Sciences. We can summarize it in a flow
chart.

DOI: 10.1093/oso/9780198825623.001.0001
16
The “What” and “Why” Questions
With the previous chapters, we completed our discussion

of the scientific method. We have seen its object of study
and limit of inquiry, so the reader should understand what
Science can and cannot address.
In these final chapters, I will try to reinforce this under-
standing by addressing questions that we, scientists, are
often asked and whose answers, sometimes, are believed to
be (incorrectly) the domain of the Natural Sciences.
Some of these questions involve terms that are borrowed
from Science, but whose content, in reality, eludes scientific
inquiry. Common examples that fall into this category
include the following:

DOI: 10.1093/oso/9780198825623.001.0001
What is gravity? What is a force? What is a particle?
As I have already addressed in this book, such questions

are not, and should not be the domain of investigation of the
Natural Sciences. The content of such questions cannot be
probed by our senses or extensions of our senses. As such,
they are beyond reach of the scientific method.
No doubt, these are questions worth asking. However,
they address the ontological nature of the subjects under study,
namely their very existence and their essence. This is what
Philosophy addresses. However, such arguments are not, by
their own nature, subject to scientific investigation.
In other words, although the above questions and many
statements that employ terms borrowed from Science may
be legitimately logical and open to philosophical discussions,
they are nevertheless not scientifically testable.
For instance, I could argue that since all material
objects in the Universe share an attraction that we
call “gravity,” irrespective of their mass, or any other
properties (“accidents,” as philosophers would call them),
the essence of gravity (what makes it fundamentally what it
is) is precisely the property of being attractive. Without
such an essential property of attraction, gravity would not be
what it is.
Of course, such an argument is open to philosophical
debate. The important point, though, is that it is unques-
tionably not a well-defined scientific statement: I cannot test
it experimentally!
While I can measure the strength and direction of
what we call gravitational force with very accurate instru-
ments, precisely because my instruments interact with the
The “What” and “Why” Questions 91
attributes of material objects (their mass), what experiment

can I think of that would measure the essence of such a
reality?
Similarly, we can measure with clocks the physical quantity
we call time. However, is there a measurement apparatus
that can measure its nature?
And I could go on and on with any material object
or phenomenon I can think of. While I can definitely
address the material world with empirical means, I cannot
use the same means to understand its existence and very
essence.
Therefore, there is no point in invoking Science to
address the “what questions” above.
These types of questions, however, are not the only ones
Science cannot address.
Another set of questions scientists are often asked
includes the following:
Why is there gravity? Why is there the Universe we observe
and not something else? Why are there animals or humans
at all? Is this the design of an intelligent being (God )?
These “why questions” address the “finality” or “motives”
of their respective subjects. As such, they cannot be
answered with scientific means.
Note, however, that these “why questions” are not the
same ones that, say, a biologist would ask regarding, e.g.,
the “scope” of a given limb in an animal, or the “reason”
why certain animals hunt in packs while others do not.
Answers to these questions can be provided within the
scientific method, since they point to the description of
causes whose ultimate effect is the survival of the individ-
uals and species.
For instance, the act of chewing food with our teeth is

related to its easy ingestion, which ultimately helps the
nutrients necessary for the survival of our body to be
digested more easily. I can then say that the survival of the
body is related to the appropriate supply of nutrients, and
hypothesize that this supply is facilitated by the chewing
of food.
So the question “Why do we have teeth?” could be equally
formulated as “What are teeth for?” And to this question,
I can provide an answer by formulating the hypothesis
above, make predictions (e.g., without teeth the food needs
to be grinded by some instrument, or I cannot swallow it),
test them, etc.
On the other hand, if I asked “Why do we want to survive
at all?” then I am asking a question that I cannot answer
scientifically. You cannot just reply, “Because we want to
propagate our species.” Otherwise, the next question would
be “But why do we want to propagate our species in the first
place?” What is the ultimate purpose of propagating our species?
Is there an experiment that I can conceive that answers
this question? The answer is no. We cannot think of any
experiment that measures the reason why we, or any other
living being on Earth for that matter, want to survive.
Answering this type of question is a task that is best left
to Teleology: a part of Philosophy that attempts to explain
phenomena by their ultimate purpose. In fact, by “teleological
argument” we oftentimes mean the one for the existence
of God from the evidence of order, and, hence, design in
Nature.
These questions, again, cannot be answered by applying
the scientific method, since they do not lend themselves
to being testable experimentally. For instance, a physicist

cannot answer the question “Why is there gravity and not
anything else?”
Science cannot answer this question simply because we
do not know, and cannot probe any other alternative, such
as a Universe without it.
Can you perform an experiment in which you eliminate
gravity in the whole Universe? Of course not.
By the same token, arguments that use the “order” in
the Universe as strong “evidence” of an Intelligent Designer
qualify as religious (or philosophical) arguments, but have no
place in Science!
In a scientific sense, the word “order” is used in many
contexts, such as in the distinction between a solid (the
“ordered state of matter”) and a liquid or gas (“disordered
states of matter”).
Does it mean that some things in the Universe are
“ordered” (like solids) and some are not? Which parts of
the Universe are ordered? What do we mean then that we
observe “order” in the Universe? How do we even quantify
such order for the whole Universe?
And, most importantly, what predictions can we extract
from that order that are experimentally testable?
Another popular argument in favor of “design” involves
the concept of “fine-tuning” of the “fundamental” physical
constants, such as the fine-structure constant, or the
charge of the electron, or the strength of the nuclear
force.
Again, without entering into worthy debate as to
whether there is a Designer, the argument of “fine-tuning”
is beyond our experimental reach.
In fact, “thought exercises” like “What happens if some of

what we call ‘fundamental’ or ‘universal’ physical constants
are changed even by a bit?” do not stand the scientific test.
There is no way for us to do that experiment and observe its
consequences!
At most, we can provide an “educated guess” as to what
would happen if one of those constants, say, the speed of
light in the vacuum, were not what it is, thus imagining
“worlds” where superluminal information is transmitted.
(Writers of science fiction novels or movies do this all the
time.)
However, excluding science fiction, we cannot change
the value of the speed of light in the vacuum, or of any
other “fundamental” constant for that matter.1 We cannot
even conceive of changing one of them while keeping the
others fixed, and test experimentally such a change.
In summary, “understanding” in Science is not the same
as “understanding” in Philosophy or Religion. In Science
we describe relations between different material objects and
their relation with us, the observers. We cannot understand
why they exist or what their ultimate purpose is.

• Science cannot answer questions about the existence and
essence of the material reality (“what questions”) and its
ultimate goal (“why questions”).
1 Note that I have added the words “in the vacuum” on purpose. The
speed of light does change in materials other than the vacuum. It is the
speed of light in the vacuum that is considered a “universal constant,”
irrespective of the relative speed of the observer and the light source.
• The “what questions” are ontological; the “why ques-

tions” are teleological.
• Therefore, they pertain to other sources of knowledge,
such as Philosophy or Religion.
• We cannot conceive of any experiment that allows an
experimental test of the above questions.
• Attempts to answer those questions only within the
boundaries of Science will undoubtedly lead to illogical
and false conclusions.
17
“Scientism”: Abusing the
Scientific Method
I started this book by quoting a sentence that I proved to be

illogical: “Science is the only way we know truth.”
Although this sentence refutes itself, it is still used in
conversations, and thus shows an alarming pattern among
some scientists, and other people who use “scientific argu-
ments” for their own agendas, or to defend their views on
Religion or Philosophy.
Such arguments tend to polarize people against each
other, rather than bringing to the table an honest and
balanced discussion.

DOI: 10.1093/oso/9780198825623.001.0001
“Scientism”: Abusing the Scientific Method 97
For example, we hear sentences like “You must be insane

or uneducated if you do not believe in . . .” (pick your favorite
scientific subject).
These sentences have two major problems.
For one, they provide ad hominem attacks, namely attacks
directed at the person(s), rather than preparing the inter-
locutors for an open discussion about the scientific subject
under review.
Secondly, they use the word “believe” as if what we are
discussing is a “divine” or “revealed” truth. Argumenta-
tions such as these are not and should not be part of any serious
discussion about Science.
They are indeed an abuse of the scientific method to
its core since, as we have seen in this book, there is no
“revealed truth” in Science. The only objective facts are
those that can be observed with our senses or extensions of
our senses. Nothing else is a fact in Science, and no amount
of “consensus” among scientists can change that, or “settle”
a scientific inquiry!
Applying the “principle” that Science is the only way we
know truth leads also to arguments that attempt to explain
the “origin of the Universe.”
For instance, we hear opinions from some scientists that
the Universe could have “emerged” from a vacuum (quan-
tum) fluctuation. Therefore, they conclude that there is no
need of a Creator, since a vacuum fluctuation can create
“something from nothing.”
It should be now evident to the reader that this is also
an abuse of Science since it utterly confuses a philosophical/
logical argument with a scientific one.
First, let us note that a quantum fluctuation is described

in Quantum Field Theory as an event that occurs at a
given time and at a specific point in space, and involves two
conjugated particles, e.g., an electron and a positron, for a
very short time.
As such, a quantum fluctuation already involves “some-
thing,” namely the existence of time, space, matter,
charge, etc.
On the other hand, as I have already mentioned, “cre-
ation ex nihilo,” namely creation from “nothing” (truly the
absence of any created thing), is properly discussed in Phi-
losophy and Religion. It is the very act of creating everything,
including time, space, matter, etc.!
In other words, in Philosophy and Religion, “nothing”
means really absence of anything, including realities such as
space and time, that in Science we must assume as a given,
to even begin our description of material phenomena.
This would be true whether the Universe were finite in
time and space, or infinite in both. Even the latter case (a
spatially infinite Universe that always existed) would not
avoid the need to question who or what has put into existence
the very realities of space, time, matter, etc.
The event of creation itself cannot be probed experimentally: it is
beyond scientific reach. You may choose to believe it or not,
but that belief is something that pertains to Religion, not
Science.
Finally, the assumption that any scientific statement
contains truth in itself may lead to the elevation of mere
hypotheses to “theories” or worse “facts.”
A typical example is the “hypothesis” of a “multiverse,”
namely the existence of many (infinite?) “universes” apart
from ours.
Despite the very appealing illustrations we see in some

magazines or news articles about “multiverses,” this con-
cept might qualify at best as a “hypothesis,” but it fails to
predict any consequences thereof that are testable. Hence, it
is again beyond the limit of scientific inquiry.
If there are other universes beyond ours and we cannot
even communicate with them, how can we measure any of
their physical properties?
And if they are connected to ours, so that, in principle, we
can send signals from ours to the others, isn’t this simply a
single, unique Universe? If so, the question should be “How
do we measure the physical presence of these different ‘parts’
of our Universe?”
Lest the reader think I am only blaming this abuse of
Science on those who promote a certain worldview, I stress
again that the abuse of the scientific method is also perpe-
trated to reach diametrically opposite conclusions.
For instance, let us consider once more the claim that
since there is “order” in the Universe, then this is “proof ” of
the existence of an “Intelligent Designer.” Such a statement,
again, may find its natural place in Religion or Philosophy,
but it has no scientific value.
To address this topic by scientific means would require
a definition, preferably mathematical, of such an “order,”
an objective description of phenomena in terms of that
“order,” and, finally, testable predictions based on it.
And even if all this could be accomplished, it would still
not answer the question of who “ordered” such a material
reality, or who has created it ex nihilo. We may answer these
questions using philosophical, logical, or religious argu-
ments, but definitely not scientific (empirical) ones.
Similarly, we hear that since the Big Bang “occurred,” it

is a direct “proof” that an act of creation happened.
As I have discussed previously, (1) we did not make
any direct observation that shows the Big Bang actually
occurred. It is a hypothesis consistent with some observations,
but not a fact in itself.
(2) Even if this event did occur, we again face the same
scientific difficulty in transitioning from the description of the
initial stages of the Universe to the creation ex nihilo of such
a Universe.
Going from a singularity in spacetime to creation of that
space, time, matter, etc. is a far cry and does not follow from
any scientific argument.
Finally, a similar argument that uses the “complexity”
of the natural world as evidence of a Creator has been
made. In the same vein as the earlier discussion about order,
“complexity” (which also in this case would require a pre-
cise mathematical definition leading to testable predictions)
cannot be adduced as a “scientific statement” to validate a
religious argument. While there could be a logical relation
between the two, it is not properly a scientific one.
For instance, “complexity” in Physics is often used to
describe situations in which a physical system has many
parts (“degrees of freedom”) interacting with each other.
This large number of elements then gives rise to some phe-
nomena that cannot be easily described as originating from
the physical properties of the single elements themselves.
In simple terms, a “complex” physical system is “more
than the sum of its parts.”
However, what I just wrote is still within the confines
of the scientific method. I can certainly describe (even
mathematically) these complex systems, make predictions

about their properties, and finally test such predictions.
How do I go from the description of these complex phys-
ical systems to who or what has created them, if empirical
means are the only ones at my disposal? The only way I
can answer this question is with philosophical or religious
arguments.
In order to do so, I can definitely make a logical or philo-
sophical step, but not a scientific one. Science has nothing to
offer here.
In conclusion, I hope I have conveyed to the reader that
abusing the scientific method beyond its empirical reach,
namely beyond what we can measure (I call this abuse
“Scientism”), can lead to non-scientific conclusions.
This does not mean those conclusions are illogical (or
irrational) or cannot otherwise be proven by a solid, logical
argument. In fact, they may have their rightful place in
Philosophy or Religion. But certainly not in Science.

• Scientism is any abuse of the scientific method for pur-
poses and in contexts that are not empirical.
• Scientism is most often used, although not always,
to reach diametrically different conclusions about
Religion.
• It is seriously damaging to our ability to reason logically.
• Rather than acknowledging the strengths and oppor-
tunities offered by all sources of knowledge, Scientism
diminishes, often ridicules, and damages their role.
• It is itself a very bad form of “religion” that should be

rejected unequivocally by laymen and practitioners alike.
• Science has no use for this sort of “religion,” and we all
would do well to guard against it.
18
Final Thoughts
It is time to recap what we have discussed in this book and

offer thoughts on some points that may still require further
discussion.
We have seen that the only object of study of the Natural
Sciences is the unique material reality that is accessible via our
senses and extensions of our senses (experimental probes).
In the scientific context, the existence of such a reality is a logi-
cal necessity: this statement cannot be probed experimentally.
As scientists, we must agree on the existence of such a
unique reality, or nothing we do makes any sense. Just imag-
ine if two scientists making experiments on the same phe-
nomenon found different results and concluded that one

DOI: 10.1093/oso/9780198825623.001.0001
must live in one reality and the other in another . . . That

would make it utterly impossible to conclude anything on
the phenomenon they are studying.
Moreover, the only knowledge Science can provide of this
material reality is the description of the interactions between
the different material objects among themselves as probed
experimentally by the observers.
Science does not and cannot answer the question of what
is the essence of such a reality (ontological question) or
why there is such reality (teleological question). These are
valid questions that belong to Philosophy or Religion, not
Science.
The scientific description attempts to create a logical synthe-
sis of a particular set (not all) of the data (objective facts)
and phenomena that we observe, by making assumptions
(statements), which we call hypotheses, to build theories.
These theories cannot be only descriptive: to have any sci-
entific value they need to be predictive of new phenomena.
Otherwise, they are, at best, hypotheses, or, at worst, fruits
of our imagination.
Theories, like most of their hypotheses, are not facts! They
are a set of statements used to build a description of specific
phenomena and predict new ones. Theories can, and will,
change as new phenomena are uncovered.
This is it! This is the whole goal of Science and its limit.
Any abuse of this methodology, such as the use of scien-
tific arguments in contexts that are not empirical, like the
proof or disproof of a Creator, is not and should not be part
of Science.
Another type of abuse might also include the recent
suggestion to change the scientific method in order to
Final Thoughts 105
accommodate certain theories that do not make experi-

mental predictions.1
In reply to this proposal, we may simply ask: How do
we objectively judge the value of such theories if there is no
material reality we can compare them with?
In fact, while we may argue (as I will do in a moment)
about the timeframe within which predictions could be
tested, we cannot debate about the necessity of such tests!
It would be as though two chess players were debating
whether the ultimate goal of the game is to checkmate the
adversary king. They may debate whether to play a short or
long game, but not on the final goal of the game itself.
Concerning the timeframe necessary to test predictions,
one of the open questions is what to make of some theories
that make predictions that cannot be tested, not just during
our lifetime, but possibly in several lifetimes.
We know, for instance, that an observation consistent
with what we call “gravitational waves” has been recently
made.
The prediction has been tested but almost 100 years
after the publication of the original idea. If we lived during
Einstein’s time, we would have thought such a prediction
so far out into the future to be almost impossible to test.
Nonetheless, we did.
How do we regard then those theories that make predic-
tions of phenomena at high energies or length scales that
are impossible to attain with present instruments?
1See, for example, Richard Dawid, String Theory and the Scientific Method
(Cambridge University Press, 2013).
It is difficult to see how we could improve our instru-

ments to reach such tiny length scales or high energies
within a century, let alone within a few decades. Do those
predictions qualify as valid?
Following the scientific methodology, the answer is yes,
but with a provision. They are valid predictions if they fol-
low logically from a set of hypotheses, and if those hypothe-
ses can also be used to describe a class of phenomena that we
have already observed.
Consider again the prediction of gravitational waves,
made after Einstein proposed his theory of General Relativ-
ity. The latter could already describe accurately the anoma-
lous precession of the perihelion of the planet Mercury.
The other predictions, such as the bending of light due to
a massive object, gravitational redshift, and many others,
including gravitational waves, were all tested subsequently.
Nonetheless, Einstein’s theory already described a class of
phenomena previously observed. It was just a matter of time
to test its other predictions and see whether they would be
corroborated (hence, expanding the reach of the theory) or
not (thus, determining its limits).
The problem arises if theories, despite not having been
experimentally tested yet, are elevated to “accepted theo-
ries” or, worse, “facts.”
There is truly no point in claiming that such theories
are “accepted” by the scientific community, since their sci-
entific value has not been really proved yet by its true,
ultimate judge: Nature.
As I have discussed in this book, the descriptive power of
a theory should not be construed as a “fact.”
To all of this let me add yet another point.
Final Thoughts 107
In some fields of the Natural Sciences, Physics in partic-

ular, there is also a tendency among theorists to put an
exaggerated emphasis on the mathematical complexity and
“intricacies” of a theory. It is as if the more “complex” the
theory is, the more “important” it must be.
This attitude completely misses the ultimate goal of a
theory: to describe phenomena and predict new ones, irrespective
of its mathematical “intricacies.”
Of course, if new types of Mathematics, as difficult as
they may appear at first, help in the prediction of new
phenomena, then they are more than welcome to enter
our toolbox of useful mathematical tools.
The Mathematics of Hilbert operators in Quantum
Mechanics or differential geometry in General Relativity
are obvious examples.
In addition, it might happen that a different, oftentimes
more “economical” (simpler) mathematical formulation of a
theory inspires new directions of study. This is all good for
the progress of Science.
However, a problem may arise if Mathematics is elevated
to the only guiding principle to discover new phenomena. This is
because doing Mathematics for its own sake, while employing
it in the Natural Sciences, may lead to logically derived
mathematical constructs that, however, have little or no
bearing on natural phenomena. In fact, by following only
logical, mathematical steps while building a theory may
even lead to completely unphysical results.
For instance, an infinite quantity is a well-defined concept
in Mathematics, but it is not physically possible in the
Natural Sciences: we cannot measure infinities because our
instruments are finite.
In general, quantities firmly grounded upon the founda-

tional axioms of Mathematics may have no correspondence
in the material world.
Therefore, while we must surely value Mathematics as
an incredibly powerful tool, it is important to guard from
overemphasizing its importance in the Natural Sciences,
whose ultimate goal is to describe the natural world, irre-
spective of what mathematical description we employ, if we
employ one at all.
In addition, a theory should be preferred to another only
if it allows describing a certain class of phenomena more easily
(Occam’s razor), or extends its reach to other phenomena that
the other theory cannot describe or predict.
If neither of these two criteria hold, namely the theory
is far too “complex” (in a mathematical sense) to allow
calculations and does not make any other prediction, in
addition to those offered by other existing theories, then
there is no reason in elevating such a theory to the rank
of “accepted theory.”
If this happens, it is typically because some individuals
have exerted some type of “pressure” on the community
to promote their own agenda.
When this happens, Science does not always gain.
I am discussing these points because both the number of
scientists and the scientific output is steadily increasing.
For this reason, it is imperative to discuss what reasonable
limits we should impose on what someone can say regard-
ing the material world. Otherwise, anything goes.
This is why the past century has seen the refereeing sys-
tem (namely anonymous peer reviews of scientists’ work)
take an increasingly important and central role in judging
Final Thoughts 109
new scientific claims, with regard to both research output

and the award of funding.
Although such a system is not infallible,2 and definitely
could be improved, it does provide some needed account-
ability on the work of scientists.
However, the discussion of how to improve it, or how to
come up with alternative forms of “checks and balances” is
still ongoing. Any suggestion is therefore welcome.
Ending on a positive note, I am convinced that if the
scientific method is applied with intellectual honesty, it will
allow us to make sense of the facts that Nature will reveal
in the process of discovery.
This is precisely because the objective reality we study is
there, irrespective of the fads of the moment, or the opinions
of a few.
By keeping this in mind, we can certainly say that
humanity will always benefit from this marvelous enter-
prise we call Science.
2 For instance, the refereeing system could be manipulated by a small
number of scientists promoting their own ideas and the ideas of their
“trusted circle,” or, worse, by some editors interested only in increasing
the popularity of their journal.
Further Readings
Pierre Duhem, Essays in the History and Philosophy of Science (Hackett

Pub. Co., 1996).
In my view, this book more than any other clearly reflects on
the limits and possibilities of Science. It was written by a well-
known physicist at the beginning of the past century, but it is
still very valid to this day. A very enjoyable read that expands
on what I have discussed in my book.
Thomas Kuhn, The Structure of Scientific Revolutions (University of
Chicago Press, 1962).
This book was written by a philosopher who trained as a physicist.
It focuses mostly on the “sociology” of Science, and how scien-
tific revolutions (changes in scientific “paradigms,” as he calls
them) have occurred historically. It is a nice compendium and
an easy read, but does not address the method as clearly and in
depth as Duhem’s book.
Karl Popper, The Logic of Scientific Discovery (Routledge, 1959).
This is a book on epistemology written by a well-known philoso-
pher of Science. One of the main tenets of Popper’s work is
that scientific methodology should always lead to falsifiability
of a scientific theory. He claimed this is because no amount of
experiments could ever “prove” a theory, but a single one is
enough to contradict it.
As I have discussed, such eventuality does not mean that the
whole theory is false. An experiment that turns out to falsify a
prediction of a theory simply sets new boundaries for the theory,
but by no means invalidates all its tenets.
Scientific theories are not mathematical conjectures, which can be
proved false by a single counter example. This is because the
goal of Natural Sciences is to describe an objective material real-
ity and predict new phenomena, whereas Mathematics extracts
conclusions that follow logically and unequivocally from its
foundational axioms, irrespective of whether they correspond
to a material reality.
About the Author
Massimiliano Di Ventra is Professor of Physics at the Uni-

versity of California, San Diego. He has published the text-
book Introduction to Nanoscale Science and Technology (Springer,
2004) for undergraduate students, and the graduate-level
textbook Electrical Transport in Nanoscale Systems (Cambridge
University Press, 2008).
Illustrations by Matteo Di Ventra (MD2 ).
Index
A God 11, 18, 37, 41, 91, 92

ants 78
Aristotle 2 H
hypothesis 11, 20, 45
B
Big Bang 100 I
Biology 6 inconsistent
data 27
C hypothesis 47
cause 17 instruments 6, 24, 33, 105
common sense 3, 4, 15 Intelligent Designer 93, 99
complexity 100
consensus 83, 86, 97 K
consistent
Karl Popper 69
data 28
hypothesis 46, 47
Copernican system 76
L
creation 18, 98, 100 Logic 2, 4, 5, 14, 15, 18, 22
logical thinking, see Logic
D
description 19, 20, 39, 42, 55, 59, 71, M
80, 83, 101 material reality 7, 10, 21, 24, 33, 99,
description limitation 42, 78 103, 105
material world 7, see material
E reality
effect 17 Mathematics 37, 55, 107
experimental limitation 42 measurement errors 26, 32, 34
microevolution 48
F
facts 28 N
faith 37 Natural Sciences ix
Falsifiability 69 Newtonian Mechanics 39, 59, 66,
70, 75
G
General Relativity 39, 49, 68, 73, O
75, 106 objective data 28
114 Index
Occam’s razor 63, 64, 75, 108 Quantum Theory, see Quantum
ontological 90, 95, 104 Mechanics
order 93, 99
R
P Religion 5, 8, 11, 94,
paradigm 84 96, 104
phenomena 7, 18, 24, 29, 37, 43, 49,
53, 59, 62, 75, 83, 92, 98, S
104, 107 Scientism xi, 101
Philosophy 1, 8, 17, 90, 92, 94, Special Relativity 39, 66, 73
96, 104 supernatural 3, 18, 37
Physics 6
Pierre Duhem x, 71, 111 T
predictions 21, 53, 58, 62, 70, 82, 93 teleological, see Teleology
Ptolemaic system 76 Teleology 92
theory 54, 62, 66, 69, 70
Q
Quantum Mechanics 9, 29, 49, 56, V
68, 79, 83, 107 verify 74

The Scientific Method Reflections From A Practitioner

Uploaded by

Copyright:

Available Formats

The Scientific Method Reflections From A Practitioner

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Scientific Method Reflections From A Practitioner

Uploaded by

Copyright:

Available Formats

THE SCIENTIFIC METHOD

I am a theoretical physicist who has been working for

convey to them both the reach and, most importantly, the

about Nature, and, at the same time, defines the boundaries

I intend it to be an easy read, and therefore I have made

than ever. In fact, by forgetting its foundations, we trans-

PS: Why the cartoons? You have heard that “a picture

1. Science without Philosophy? 1

Further Readings 111

A first question to settle before embarking further into the

Ultimate Questions of Life (Bantam, 2010).

The Scientific Method. Massimiliano Di Ventra, Oxford University Press (2018).

Let us, for instance, analyze the first statement: “Science

We reach that conclusion by implicitly using a form of

I wanted to stress this point, because, in conversations

Takeaways from this chapter

All scientific disciplines tend to specialize in particular

The Scientific Method. Massimiliano Di Ventra, Oxford University Press (2018).

naked eyes. Telescopes have been built to precisely expand

can address. Of course, we can, and do deal with it within

I have encountered people that would deny the existence

same physical property must be the same irrespective of the

Note that this limit of inquiry is trespassed by the previous

In that example I concluded that if I found a road wet,

possible phenomenon that could have produced the observed

Takeaways from this chapter

Once we have settled on the issue that there is such a thing

The Scientific Method. Massimiliano Di Ventra, Oxford University Press (2018).

several important first principles or self-evident truths (or laws of

The law of identity

The law of non-contradiction

The law of the excluded middle

1 Metaphysics 4.4, W.D. Ross (trans.), GBWW 8, 525–6.

The attentive reader would immediately recognize that

This is, indeed, an application of the “law of excluded

The law of cause and effect

There are actually two ways to look at this principle: one

phenomena. In fact, those causes could be anything in the

In simple terms: there is no way to experimentally test, with

Yet more: the atoms of the child’s foot, collectively,

more “elementary cause” behind the momentum imparted

Hopefully, all this discussion should have elucidated even

Takeaways from this chapter

3 The majority of—and in a very strict, formal sense, all—statements

in Mathematics are like this. Mathematical statements only need to

• This description, however, needs to lead to testable pre-

We have now established that the object of study of the

The Scientific Method. Massimiliano Di Ventra, Oxford University Press (2018).

However, what part of the material reality do we actually

Despite what we may sometimes hear, unless we are

be slightly different from the others (say, one kilometer

but doable), or we realize that one of the two (or both)

Objective data or facts

How many times should a measurement be performed?

1The “error analysis” I have presented here is very simplistic. It is

The data are then declared objective because another

mation we collect of the natural world (the objective

Takeaways from this chapter

• Data about a given phenomenon are objective when