Pragmatic Information in Quantum Mechanics

arXiv:1512.05651v2 [physics.hist-ph] 7 Apr 2016

Juan G. Roederer
Geophysical Institute, University of Alaska Fairbanks
Fairbanks AK, 99775, USA
April 8, 2016

Abstract : An objective definition of pragmatic information and the consid-

eration of recent results about information-processing in the human brain can
help overcome some traditional difficulties with the interpretation of quantum
mechanics. Rather than attempting to define information ab initio, I introduce
the concept of interaction between material bodies as a primary concept. Two
distinct categories can be identified: 1) Interactions which can always be re-
duced to a superposition of physical interactions (forces) between elementary
constituents; 2) Interactions between complex bodies which cannot be reduced
to a superposition of interactions between parts, and in which patterns and
forms (in space and/or time) play the determining role. Pragmatic information
is then defined as the correspondence between a given pattern and the ensuing
pattern-specific change. I will show that pragmatic information is a biological
concept that plays no active role in the purely physical domainit only does
so when a living organism intervenes. The consequences for the foundations
of both classical and quantum physics are important and will be discussed in
detail. Since by its very definition pragmatic informationthe one our brain
uses to represent, think about and react to the outside worldcannot operate
in the quantum domain, it is advisable to refrain from using it in our mental
representation of what is happening inside a quantum system. Although the
theoretical framework developed for quantum mechanics handles mathematical
entities specifically attributed to a quantum system, the only truly pragmatic
information it can provide relates to macroscopic effects on the environment
(natural, or artificial as in a measurement instrument) with which the system
interacts.

1 What is Pragmatic Information?

We live in the Information Era. It is quite understandable that a term of such
every-day usage as information remains largely undefined when used in the
general scientific literature, despite the fact that philosophers, mathematicians,
Permanent address: Frasier Meadows, 4825 Sioux Drive, Apt 102, Boulder, Colorado

80303, USA. URL: www.gi.alaska.edu/Roederer

1
linguists, engineers, biologists and writers use this common term with quite
different and distinct meanings. Physicists are accustomed to working with
the mathematically well-founded concepts of Shannon (or statistical) and algo-
rithmic information. These terms are used to designate objective, quantitative
expressions of the amount of information in a message; the gain of information
when alternatives are resolved; the degree of uncertainty; the quality of infor-
mation; the minimum number of binary steps to identify or describe something;
the maximum number of bits that may be processed in one location; the total
number of bits available in the Universe; etc. However, whenever physicists
use the term information in statements in a qualitative, descriptive way such
as the field carries information about its sources; information cannot travel
faster than light; information about the microstate of a gas; the detector
provides information about the radiation background; information deposited
in (or extracted from) the environment; etc., they are really talking about
pragmatic information.
In effect, it is mostly the biologists, particularly geneticists and neurosci-
entists, who make an explicit distinction between Shannon or statistical infor-
mation (detached from purpose or effect), algorithmic information (minimum
number of binary steps to define something, or the numerical values of physi-
cal variables) and pragmatic information (linking purpose and effect) (e.g., see
Kuppers 1990). Of these three classes of information the last one is usually the
most relevant in biology; indeed, the notion of quantity of information is often
of secondary importance: what counts is what information ultimately does, not
how many bits are involved.
As we argue in Roederer (2005), pragmatic information is a purely biolog-
ical concept. This seems quite difficult to accept for physicists, not the least
because many agree with John Wheelers dictum that every physical quan-
tity derives its ultimate significance from bits, binary yes-or-no indications
(Wheeler 1989). Yet let me emphasize the trivial facts that physics is the prod-
uct of information-processing by brains, based on interactions of systems chosen
and prepared by brains, measured with artifacts (instruments) created by brains,
using algorithms (mathematics and models) developed by brains and following
plans, purposes and quantitative predictions made by brains. Perhaps the prin-
cipal obstacle in an effort to persuade physicists to accept the idea of information
being a purely biological concept, is the perceived lack of a widely accepted def-
inition of pragmatic information that is truly objective, i.e., unrelated to any
human use or practice.1
Philosophers have struggled for a long time with an ab initio definition of
information. Physicists often consider information in its general sense to be a
statement that resolves a set of alternatives, i.e., which resolves uncertainty.
Others consider data to be the essence of information. Most visualize infor-
mation as a quantity expressed in number of bits. Rather than attempting to
define information ab initio, I find it more appropriate to start with interaction
1 From its beginning, traditional information theory (Shannon and Weaver 1949) deliber-

ately refrained from giving a universal and objective definition of the concept of information
per se.

2
between bodies as the primary concept or epistemological primitive (Roederer
1978, 2005), and from there derive the concept of information. We can iden-
tify two distinct groups. Category 1) interactions can always be reduced to a
linear superposition of mutual physical interactions (i.e., forces) between the
interacting bodies elementary constituents, in which energy transfer between
the interacting parts plays a fundamental role. Category 2) interactions cannot
be expressed as a superposition of elementary interactions, and it is patterns
and forms (in space and/or time) that play the determining role on whether or
not an interaction is to take place; the required energy must be provided from
outside through a specific complex interaction mechanism.
Examples of category 1), which we call force-driven interactions, are all the
physical interactions between elementary particles, wave fields, nuclei, atoms,
molecules, parcels of fluid, complex solid bodies and networks, planets and stars.
Ultimately, they all originate in the four basic interactions between fundamental
particles (electromagnetic, gravitational, strong and weak). It is the fundamen-
tal property of reducibility of force-driven interactions, that allows physics to
work with approximate models of the complex reality outside to make quan-
titative predictions or retrodictions about the time-evolution of a given physical
system.
The simplest case of an interaction of category 2) is any arrangement in
which the presence of a specific pattern in a complex system S (sometimes
called the sender or the source) leads to a causal, macroscopic and univocal
change in another complex system R (the recipient), a change that would not
happen (or just occur by chance) in the absence of the particular pattern at the
source. Typical examples range from effects on their respective chemical envi-
ronments of the one-dimensional pattern of bases in the RNA molecule or the
three-dimensional shape of a folded protein; the light patterns detected by an
insect and the resulting shape of its flight orbit around a light source; the pat-
terns of neural electrical impulses in one region of the brain triggering impulses
in another; to the perceived print patterns changing the state of knowledge in
a readers brain. It is important to point out that information-driven interac-
tions all require a complex interaction mechanism with a reset function (often
considered part of one of the interacting bodies), and which ultimately provides
the energy required to effect the specific change. Although we call them in-
teractions, they are causally unidirectional, from source to recipient. However,
the designation inter -action is justified in the sense that, to occur, they require
some predesigned or evolutionary match (sometimes called understanding)
between source and recipient.
Pragmatic information is then defined as that which represents the univo-
cal correspondence pattern change; it is the reason why we call Category 2)
information-driven interactions (Roederer 2003).2 By univocal we mean that
the interaction process is deterministic and must yield identical results when
2 This may appear to some readers as a circular definition. It is not: information-driven

is just a convenient name for this category; we could have called them pattern-driven inter-
actions.

3
repeated under similar conditions of preparation3. This in turn means that the
triggering pattern must be stable during some finite time and be amenable to
be copied. When an information-driven interaction has occurred, we say that
pragmatic information was transferred from the source S (where the pattern
resides) to the recipient R (where the pattern-specific change occurs). We wish
to emphasize that in this definition, the concept of pattern refers to a physi-
cal/topological property in space and/or instants of time;4 it includes, but is not
limited to, symbols to which one can assign syntactic and semantic dimensions
(e.g., see Kuppers, 1990, Chapter 3).
Note that in the preceding we do not imply that information resides in the
patternsthe concept of pragmatic information is one of relationship between
patterns and changes, mediated by some interaction mechanism. A pattern all
by itself has no meaning or function. On occasion one may say that informa-
tion is encoded in the pattern at the source. The effected change can itself
be a pattern. Finally, there is no such thing as a numerical measure of prag-
matic information. Pragmatic information cannot be quantifiedit represents
a correspondence which either exists or not, or works as intended or not, but
it cannot be assigned a magnitude. There are cases, of course, where a given
pattern and/or the change can be expressed in numbers or bits by an external
agent, but that number is not the pragmatic information involved.
There are only three fundamental processes through which mechanisms of
information-driven interactions can emerge (Roederer 2005), involving processes
at three vastly different time scales: 1) Darwinian evolution; 2) adaptation
or neural learning; 3) as the result of human reasoning and long-term plan-
ning. In other words, they all involve living matter indeed, information-driven
interactions represent the defining property of life (Roederer, 2004). In this
overall evolution, we note a gradual increase of complexity of the interacting
systems and related mechanisms involved: from simple chemical reactions
between biomolecules, to life-sustaining circulation systems in plants, to the
ultra-complex interaction chains in the human brain.
Any information-driven interaction between inanimate complex systems must
ultimately be life-generated or -designed, requiring at some stage goal-directed
actions by a living system. Examples are the physical effects that a beaver dam
has on water flow; mechanical effects on the environment of a tool used by a
corvid; flight path control by an autopilot, etc. As a more explicit example,
consider an electromagnetic or sound wave emitted by a meteorological light-
ening discharge, which does not represent any information-driven interaction:
it is generated and propagates through physical processes in which information
plays no role. But waves emitted by an electric discharge in the laboratory may
be part of an overall artificial information-driven interaction mechanism if they
are part of a device created by a human mind with the intention of having the
discharge cause a desired change somewhere else, such as a record on a tape or
a mental event in an observer.
3 We shall ignore at this stage considerations of matching errors, fluctuations. etc.
4 In philosophy there is a large literature on semiotics, the study of signs, their meaning
and effects, first developed by C. S. Peirce in the late 19th century.

4
2 Information and Classical Physics
All information-driven interactions, whether purely biological or in a human-
controlled scientific experiment, affect the normal non-biological course of
physical and chemical events. Information-driven interactions all involve com-
plex systems in the classical domain, with time-sequences which fulfill the dic-
tates of causality, locality, special relativity and thermodynamics. It is impor-
tant to emphasize that information-driven interactions indeed do function on
the basis of force-driven interactions between their componentswhat makes
them different is how these purely physical components are put together in the
interaction mechanism (the informational architecture of Walker et al., 2016).
The important question of when a force-driven interaction becomes information-
driven will be briefly addressed later.
The concept of information does not appear as an active, controlling agent
in purely physical interaction processes in the Universe; it only appears there
when a life system in general, or an observer in particular, intervene (see Roed-
erer 2005, Chapter 5). In other words, the world out there works without
information-processinguntil a living system intervenes and changes the phys-
ical course. The above-mentioned beaver dam is an example. In physics, when
we state that a system of mass points follows a path of least action, we do
not mean that the system possesses the necessary information to choose a path
of minimum-action from among infinite possibilities, but that it is we humans
who have discovered how systems of mass points evolve and who developed a
mathematical method applicable to all to predict or retrodict their motions.
Similar arguments can be made when we describe black holes as swallowing
information, or decoherence as carrying away information on a quantum sys-
tem. In summary, the physical universe does not obey physical lawsit is
us the observers who make the laws based on systematic observations (ulti-
mately, specific changes in our brainssee section 3), of the changes left in our
instruments by their exposure to environmental events. And it is us who are
able to devise a mathematical framework that entices us to imagine (see sec-
tion 3) processes like black holes being an information-processing and -erasing
system, or an electron being at two different places at the same time.
Further examples are found in the association between entropy and infor-
mation, which arises from the particular way we scientists describe, analyze
and manipulate nature, for instance, by counting molecules in a pre-parceled
phase space; coarse-graining (averaging over preconceived domains); looking for
regularities vs. disorder; quantifying fluctuations; extracting mechanical work
based on observed patterns in the system; or mentally tagging molecules ac-
cording to their initial states. In this latter context, let us discuss a concrete
example, which also will be helpful later in a discussion of information in the
quantum domain. We turn to Gibbs paradox (for details see Roederer (2005),
section 5.5). Consider two vessels A and B of equal volume, joined by a tube
with a closed valve, thermally isolated and filled with the same gas at the same
pressure and temperature. If we open the valve, nothing will happen ther-
modynamically, but at the microscopic level, we can picture in our mind the

5
molecules of A expanding into vessel B and the molecules of B expanding into
A. Each process would represent an adiabatic expansion with an increase of the
entropy by kN ln 2 (where k is Boltzmanns constant and N is the number
of molecules in A or B), so there should be a total increase of the entropy of
the system by S = 2kN ln 2. This of course is absurd and represents Gibbs
paradox. In most textbooks it is (somewhat lamely) explained away by saying
that the formulas used here are indeed correct, but apply only if the gases in
the two vessels are physically distinguishable 5 (different gases, or just specimens
of opposite chirality).
So what is wrong with the conclusions of the above thought experiment?
By invoking the concept of pragmatic information when we say that we can
picture in our mind the molecules from vessel A doing this or that. . . , we are
labeling them so as to make them different from the others, even if in reality
they are not. In other words, we are assuming that on each molecule there
is a pre-established pattern (which triggers in our brain the neural correlates
this molecule is from A, that molecule is from B). However, this pattern and
any extractable information from it do not exist we just have forced them into
our mental image! The same argumentation, invoking the concept of pragmatic
information where there is none in reality, can be made regarding Maxwells De-
mon paradox. Notice carefully that whenever we invoke pragmatic information
in our mind, as in a Gedankenexperiment (i.e., establishing some imagined cor-
relation between a pattern and a change governed by some imagined interaction
mechanism), we are dealing with a thermodynamically open system, even if in
reality it is not.
Notice that in all examples above, the interactions involved are force-driven;
but whenever we use the term information in their description we really mean
pragmatic information for us, the observers. And when it is us who delib-
erately set the initial conditions of a classical mechanical system (or prepare a
quantum system), we are converting it into an information-driven system with
a given purpose (to achieve a change that would not happen naturally with-
out our intervention). All laboratory experiments, whether a simple classroom
demonstration or a sophisticated table-top quantum experiment, fall into this
category.
It will be useful for our discussion of quantum systems to identify common
features in a classical measurement process. First of all, note that it necessarily
involves force-driven interactions, but they are controlled by a human being or
a preplanned artifact. For instance, when you measure the size of an object
with a caliper, the instrument interacts elastically with the object; if you mea-
sure it with a ruler, you need to submerge everything in a bath of photons
whose scattering or reflection is what your optical system uses to extract the
wanted information on objectapparatus interaction. Note that in the latter
example, the environment has become an integral part of the instrument (this
will be important for the discussion of quantum measurements). In summary,
5 Notice the link with the entropy increase/decrease per bit of Shannon information (see

Roederer, 2005): = S/(2N ) = k ln 2.

6
the measurement process represents an information-driven interaction between
the object or system to be measured and a measurement apparatus or device
(which contains a reference which we call the unit). The whole process defines
the magnitude (the observable in quantum mechanics) that is being measured.6
And it has an observer-related purpose: that of changing his/her state of knowl-
edge in a very specific way. It is clear then, that even in classical physics, it
is impossible to detach the measurement process from the observer (or his/her
measurement artifacts). In summary, the fundamental interaction stages in any
measurement process are: [patterns from an object] [change in the appara-
tus], and [pattern of change in the apparatus] [change in the state of the brain
of the observer], who thus has extracted pragmatic information from the object.
Of course, in most cases that information can be represented as a number (value
of an observable) and linked to Shannon or algorithmic information.
And here we should turn to another fundamental, inextricable link of the
concept of information to biology, by pointing out that all the operations men-
tioned above are ultimately related to how the human brain of the observer
creates internal mental images, plans, makes decisions, and reacts to and pro-
cesses external sensory information. As already mentioned above, these opera-
tions ultimately have the purpose of changing the neural cognitive state from
not-knowing to knowing, a transition which I venture to describe as the
reduction of an initial brain state involving multiple expectancies to one of pos-
sible basis states, where each basis state represents the mental image of only
one possible outcome of the expected alternatives (we might even call them
preferred mental states). Note that as a corollary, the concept of probability,
usually defined mathematically as the result (limit) of a specific physical oper-
ation (e.g., a series of measurements under equal conditions like tossing of dice,
playing roulette), has a very subjective foundation in human brain operation.

3 Information and Brain Function

Since the beginning of quantum mechanics, physicists have been arguing about
whether the observer and his/her state of knowledge, even consciousness, play
an active role in the quantum measurement process. However, they did not
have the benefit of knowing what is known today about the neurobiological
mechanisms that control human brain function.
Recent studies with functional magnetic resonance imaging, positron emis-
sion tomography, diffusion tensor imaging and, at the neural network level,
multi-microelectrode recordings, are confirming a hypothesis long in use by
neurophysicists and computer scientists, namely that the information being pro-
cessed in the brain is encoded in real-time as a task-specific spatio-temporal dis-
tribution of neural activity in certain regions of the cerebral cortex (e.g., Tononi
and Koch 2008). If, because of neural interconnectivity, a certain specific pat-
6 e.g., weight of an object is the number we obtain when we do this and this and that,

under these or those conditions. An important requirement in this is that different equivalent
recipes yield the same number.

7
tern of neural activity distribution in one area triggers a specific distribution
in another, and does so in univocal way (within limits), we are in presence of
an information-based interaction between two cerebral regions; the pragmatic
information involved represents that specific relationship, and we usually say
that information has been transferred from one cerebral region to another.
When a scientist makes a measurement, the pragmatic information involved
represents the correlation between an external pattern (e.g., the location in a
reference space of a rigid body at a given time, the position of the dial in an
instrument, the dots on cast dice, the color change of a solution) with a specific
spatio-temporal pattern of neural activity in the prefrontal lobes, corresponding
to the knowledge its this particular state and not any other possible one.
The actual information-processing mechanisms in the brain linking one neural
distribution with another are controlled by the actual synaptic wiring, which
in certain regions, especially the hippocampus, has the ability of undergoing
specific changes as a function of use (plasticity)the physiological expression
of stably stored pragmatic information or long-term memory.
Modern neurobiology has an answer to the common question: When does
a specific distribution of neural firings actually become a mental image? This
neural activity distribution does not become anythingit is the image!7 In
summary, the dynamic spatio-temporal distribution of neural impulses and the
quasi-static spatial distribution of synapses and their efficiencies together are
the physical realization of the global state of a functioning brain at any instant of
time. Another way of expressing this: pragmatic information is encoded in the
brain dynamically in short-term patterns of neural impulses and statically in the
long-term patterns of synaptic architecture. Given the number of interacting
elements ( 1012 neurons and 1014 1015 synapses in the human brain)
and the discontinuous nature of activity distribution, there is little hope that a
quantitative mathematical theory of integral brain function would be developed
in the foreseeable future.
Quite generally, animal brains handle pragmatic information in sequences
of information-driven interactions in which one specific spatio-temporal pattern
of neural activity is mapped or transformed into another neural patternin
its most basic form, from a physically triggered sensory or interoceptive stim-
ulation pattern to a neural output pattern controlling muscle and gland fibers,
thus governing the animals integral behavior. These processes may change the
interconnectivity (synaptic architecture) of participating neural networks (the
learning process), leading to long-term storage of information. A memory recall
consists of the replay of the original neural activity distribution that had led to
the synaptic changes during memory storage; the most important type is the
associative recall, in which the replay is triggered by a cue embedded in the on-
going neural activity distribution (for examples, see Roederer, 2005). Expressed
in terms of pragmatic information: in the act of remembering or imagining a
7 Just as the idea of information being a purely biological concept is unpalatable to

many physicists, the idea of a mental image, even consciousness itself, being nothing but
a very specific, unique spatio-temporal distribution of neural activity is unpalatable to many
psychologists, philosophers and theologians.

8
certain object, information on that object stored in the synaptic architecture
of the brain is retro-transferred in the form of neural impulse patterns, mostly
via subcortical networks, to the visual cortex and/or other pertinent sensory
areas, where it triggers neural activity specific to the actual sensory perception
of the object. Neuroscientists call the specific microscopic distribution of neural
activity responsible for any subjective experience a neural correlate.
Let us point out at this stage that in the definition of pragmatic information,
it is often the case that different source patterns can lead to the same change
in the recipient (e.g., different shapes, sizes and colors of an apple still trigger
the neural pattern that defines the concept or image of apple). Likewise,
the same source pattern can lead to different effects, depending on collateral
information-processing activity of the system.
The human brain can recall stored information at will as images or represen-
tations, manipulate them, discover overlooked correlations8, and re-store mod-
ified or amended versions thereof, without any concurrent external or somatic
input it can go off line (Bickerton 1995). This is information generation
par excellence and represents the human thinking process (e.g., Roederer 1978,
2005). Internally triggered human brain images, however abstract, are snippets
(expressed as many different but unique patterns of neural activity in specific
regions of the cortex) derived from stored information acquired in earlier sen-
sory or mental events, and pieced together in different ways under some central
control (the main program) linked to human self-consciousness.
Whenever a physicist conceives or thinks about the model of a physical
system or physical process, whether classical, relativistic or quantum, whether
one-dimensional or multi-dimensional, his/her brain triggers, transforms and
mutually correlates very specific and unique distributions of neural impulses.
The fact that the brain is an eminently classical information-processing device9
that evolved, and is continuously being trained through information-driven in-
teractions with the classical macroscopic world, is very germane to how we can
imagine, describe and understand the behavior of any systems, either quantum,
classical, molecular or biological. This even applies to mathematics, most no-
tably probability theory (see Roederer, 2005, section 1.6). For instance, given a
set of mental images of possible outcomes of a future quantum measurement, all
may be subjectively viewed as equiprobable by an unbiased observer. Only after
personal experience with multiple measurements under identical conditions, or
through information from others who already have undertaken this task, can
the observer develop an objective sense of traditional probability.
8 In animals, the time interval within which causal correlations can be established (trace
conditioning) is of the order of tens of seconds and decreases rapidly if other stimuli are present
(e.g., Han et al. 2003); in humans it extends over the long-term past and the long-term future
(for a brief review of human vs. subhuman intelligence, see for instance Balter 2010). Most
importantly for our discussion, this leads to the conscious awareness of the past, present and
future, and the quantitative conception of time.
9 Quantum decoherence times in the brain cells would be ten or more orders of magnitude

shorter than the minimum time required for any cognitive operation (e.g., Schlosshauer 2008).

9
4 Information and Quantum Mechanics:
Bits from Its
In the preceding sections we have identified two categories of interactions be-
tween bodies or systems in the universe as we know it: force-driven and information-
driven. The first category is assumed to be operating in the entire spatial-
temporal domain, from the Planck scale up, between elementary particles,
atoms, molecules, complex condensed bodies, networks and systems of bod-
ies. The second category leads to the definition of pragmatic information as
representing a physical, causal and univocal correspondence between a pattern
and a specific macroscopic change elicited by some complex interaction mech-
anism. By this very definition, the domain of validity of information-driven
interactions, and therefore of the concept of pragmatic information per se, is
limited to the classical macroscopic domain. The reason is that, as we shall dis-
cuss extensively below, in the subatomic quantum domain spatial or temporal
patterns cannot always be defined or identified in a univocal, stable or causal
way. This in turn is related to fundamental and exclusive properties of quan-
tum systems, which we shall examine now taking into account the definition of
pragmatic information given in section 1.

4.1 Revisiting some relevant quantum facts

For the purpose of exploring the role of pragmatic information in quantum me-
chanics it is opportune to present, in greatly simplified fashion, some experiment-
based quantum facts in a nutshell.10 To emphasize that the peculiar behavior
of a quantum system is not the result of some mathematical properties of linear
algebra and differential equations but, rather, representative of physical facts
happening out there, we shall deliberately refrain from referring to abstract
postulates, theorems and properties of Hilbert spaces with which most textbooks
introduce the subject.
Given a single quantum system that has been physically prepared in a cer-
tain way, it is impossible in principle to determine or verify its particular state
by a measurement. Indeed, having many similar quantum systems prepared in
identical ways and subjecting each one to identical measurements, one will ob-
tain a collection of measurement values from among a common (often discrete)
set that only depends on the setup and the instrument, called the eigenvalues
of the observable in question (as mentioned in the previous section, the observ-
able itself being defined by the instrument). The distribution of probabilities of
occurrence of eigenvalues (including the latters mean value and standard devi-
ation) is what characterizes the common state of each quantum system of the
set at the time of measurement (Borns Rule). While it is impossible to predict
the outcome of the measurement of any one system, the set of probabilities
of obtaining the eigenvalues is deterministic in the sense that it depends only
10 An excellent discussion of the principles of quantum mechanics as relevant to this article

is given in Chapter 2 of Schlosshauer (2008).

10
on the preparation and history of the quantum system and on the apparatus
used.11 This is not all: when a single quantum system is measured and one of
the possible eigenvalues of the corresponding observable is obtained, that quan-
tum system emerges in a special state such that a repeated measurement of
the exactly same kind immediately thereafter will always yield the same eigen-
value.12 Such a state is therefore called an eigenstate (also basis state). One
says that the initial state of the quantum system has collapsed into, or been
reduced to, an eigenstate as the result of the first intervention. In other words,
after the measurement the quantum system behaves classically with respect
to the variable in question.
Sometimes, the same apparatus can be used in different configurations (e.g.,
rotation of a Stern-Gerlach set-up or a polarization filter); experiments show
that each configuration may have a different set of basis states. In other cases
there are no alternatives for the same instrumental setup (for instance, the
which-path observable in a Mach-Zehnder interferometer; the wavelength of
a spectral emission; the decay products of a particle); the eigenstates are then
called preferred basis states. On the other hand, one single quantum system
may have more than one observable to measure (several degrees of freedom like
the position, momentum and spin vectors of a particle); in that case, the mea-
surement results of one observable may or may not affect (be correlated with) the
distribution of measurement results of another. This means that having many
multi-degree of freedom quantum systems prepared in identical ways, making
the measurement of one specific observable on some and measurement of an-
other observable on the rest, the statistical distributions of results in both series
of measurements may or may not be correlated (non-commuting or commuting
variables, respectively).
It is when non-commuting observables enter the picture that Plancks univer-
sal constant h appears and delimits the size of the quantum domain. Histori-
cally, it was the position and the momentum of a particles which led Heisenberg
to his uncertainty relationship between their standard deviations in measure-
ments under similar conditions (xp h/2). For the sake of completeness
of this summary, we must also mention that Schrodingers equation is intro-
duced when the time evolution of a quantum system is to be described (which,
however, will not be dealt with in this article).
All these experimental facts indeed allow the development of linear algebra
algorithms in which general states of a quantum system (those which, when
measured under equal conditions, give different results from among a set of
eigenvalues) are represented by (unit) vectors in a Hilbert space; interactions
such as measurements and transformations (e.g., mutual interactions, time evo-
11 There are classical systems that do behave this wayfor instance a pinball machine or a

roulette. They are the so-called deterministic chaotic systems, which during the preparation
process suffer unpredictable and unavoidable infinitesimal changes which then lead to finite
differences with which the system comes out of the preparation process. Table-top experiments
with quantum systems involving single particles (e.g., Mach-Zehnder interferometry, electron
diffraction) convincingly show that no such final differentiation happens during preparation.
12 We exclude so-called destructive measurements, in which the quantum system disappears

(e.g., is absorbed).

11
lution) are represented by specific operators. Each observable defines a sub-
space whose axes represent its eigenstates or basis states when measured with
the observable-defining apparatus in a specific configuration, and the squares of
the projections of the state vector onto the axes represent the probabilities of
occurrence of the corresponding eigenvalues. A general quantum state (called a
pure or superposed state) is thus represented as a linear superposition of basis
states; a measurement is represented by the rotation of the state vector onto
one of the axes.13 Finally, the functional expressions (e.g., the Hamiltonians)
governing the dynamic evolution of the obervables of a quantum system are
equivalent, in many cases identical, to the corresponding functional expressions
for an equivalent classical system.
Here comes our first real encounter with the concept of information. The
mathematical formalism developed for quantum systems in correspondence with
the formalism of classical mechanics tempts us to picture in our mind a super-
posed state as the quantum system being in different eigenstates at the same
time. As stated before, for a single quantum system it is impossible to deter-
mine through measurement whether it is or not in such a superposition, and if
it is, what that superposition really is. In other words: it is impossible to ex-
tract pragmatic information from a single quantum system. This is no different
than wanting to imagine in the above-mentioned Gibbs paradox the molecules
from one container moving into the other preserving a distinguishability as if
they were carrying labels from which to extract pragmatic informationwhen
in reality there is none.
Let us consider an ensemble of N equally prepared, identical quantum sys-
tems. We measure one given variable on each one and compile the results, find-
ing a whole set of different outcomes in a certain statistical distribution. This
tells us that the preparation process had produced an ensemble of quantum sys-
tems, each one in the same coherent superposed state. Now after the collective
measurement was made, we are left with a radically different kind of ensemble,
namely a mixture of N systems in which each component is now collapsed into
a fixed, repetitively measurable eigenstate.14 Pulling out just one component
at random, we would never be able to predict what its state is. But after the
initial measurements we could have labeled each element with the pertinent re-
sult (knowing that a repeated measurement would yield the same value) and
obtain a mixture of systems that would behave classically with respect to the
observable in question, because now we can extract pragmatic information from
it (e.g., we could create stable patterns and establish univocal correspondences
with changes or new patterns elsewhere).
Next, we consider an ensemble of several distinct but initially non-interacting
quantum systems in superposed states. Each one is described by its own separate
state vector in its own Hilbert space (corresponding to the observable(s) chosen).
If measured individually, we obtain individual distributions of results for each
13 Complex numbers are used for the state vector components to accommodate the possibility

of having two or more superposed states with the same probability distribution of measurement
results, but which behave differently in interactions (interference phenomena).
14 Remember that in all this we are ignoring possible changes as a function of time.

12
component. Let us now, instead, bring these individual quantum systems into
mutual interaction at the quantum level (i.e., shielded from interactions with
macroscopic bodies), take them away from each other, and only then make the
measurements. Like before, we cannot predict what the individual outcomes
will be, but once obtained, experiment tells us that they appear correlated
regardless how far apart they were taken. This means that once a set of quantum
systems in superposed states interact, they will lose their independence: the
ensemble must be treated as a single whole, described by just one global state.
These component parts are then said to be entangled. The actual results of
measurements on entangled components (unpredictable in themselves, except
statistically) will be correlated, no matter how far away in space and time they
are located from each other after the mutual interaction (there is no equivalent
example of this in classical physics). It is often stated: once the measurement
has been made on one system, the other one will automatically collapse into
a conjugate basis state. Yet there is no information transfer involved in this
process, no superluminous communication between the two! All we find (at
the macroscopic level) is that the measurement results will be correlatedbut
correlation does not imply causation!
It is also an experimental fact that not only man-made laboratory mea-
surements but any interaction with the natural macroscopic environment will
eventually break up the global state of an ensemble of entangled quantum sys-
tems into independent, ultimately collapsed states of its individual components.
This process is called decoherence (e.g., Schlosshauer, 2008). As a matter of
fact, it is impossible to completely shield a quantum system from unpredictable
macroscopic influences of the environment, and superposed quantum states are
difficult to maintain in the laboratory.
Entanglement allows us to take a better look at the quantum measurement
process, in which a quantum system is deliberately made to interact with a
macroscopic device. This device is constructed in such a way that in the special
case when the quantum system is in an eigenstate of the observable which the
apparatus is supposed to measure, the initial perturbation propagates through
the instrument and is amplified to give an observable macroscopic effect (e.g.,
the particular position of a dial) that depends on the initial eigenstate. When
the quantum system to be measured is now in a superposed state, it gets entan-
gled with the quantum end of the apparatusbut as subsequent interactions
propagate through the instrument, inevitable entanglement with the environ-
ment will eventually cause the single quantum state of the total system (mea-
sured system plus instrument) to break up into mutually correlated basis states:
(i) the instrument signals a specific macroscopic change, and (ii) the original
quantum system emerges in a corresponding eigenstate.15 Because of the hu-
man intention involved in the construction or use of a measurement apparatus,
does a quantum measurement process represent an information-driven interac-
tion? There was no initial pattern to be univocally mapped into a macroscopic
15 The Schrodinger Cat paradox refers to what would happen if there was no decoherence

in this process.

13
change (like the position of a pointer)the only information-driven interaction
occurs between the classical end of the instrument and the observer (sensory
effect of the macroscopic change in the instrument on the observers brain or a
human-designed recording device).
When an ensemble of quantum systems has decohered, it will behave clas-
sically, at least with respect to the observables whose states have collapsed,
furnishing the same results in immediately successive, identical measurements.
Only then can we define patterns based on measurement outcomes for information-
driven interactionssuch interactions cannot occur at the non-decohered quan-
tum level. We can emphasize again: pragmatic information cannot exist and
operate in, or be extracted from, a pure quantum domain. A logical consequence
is that a single quantum system in a superposed state cannot be copied (the
so-called no-cloning theorem). Well come to this again in the next section.
It is an experimental fact that macroscopic bodies are in decohered and re-
duced states, particularly of the position variable, but so are also the atomic nu-
clei of organic molecules.16 The latter behave classically and can carry, transfer
or respond to pragmatic information: a spatial pattern in some organic molecule
(e.g., RNA) triggering a specific change in some other molecular system (e.g.,
formation of a protein), with some complex molecular organelle (e.g., a ribo-
some) responsible for the interaction mechanism, is possible! Herein lies the
essence of the evolutionary emergence of life systems (Roederer, 2005).
Let us now walk through some well-known simple examples, usually taught
at the beginning of a course on quantum mechanics or quantum computing, and
examine them from the point of view of pragmatic informationpointing out
where the latter plays a role and where it cant.

4.2 Single qubits

Consider the measurement of a qubit (elementary quantum system with only
two possible eigenstates, like a 1/2 spin silver atom, a photon in a two-path
Mach-Zehnder interferometer). The qubit interacts with the apparatus at its
quantum end (e.g., the atoms to be ionized in a particle detector), and after
the measurement the classical end exhibits a macroscopic change (e.g., voltage
pulse, a blip on a luminescent screen, position of a pointer, a living or dead
cat). It is the physical structure of the device enabling the occurrence of such
macroscopic change that defines the observable in question (although in this
case it is better to say that, given the qubit, its two preferred basis states
determine the instrument that must be used). The observer per se is irrelevant
during the measurement itself, except that we must not forget that it was a
human being who set up the measurement (hence decided what observable to
measure), who selected and prepared the qubit to be measured, and whose brain
ultimately expects to receive an image, the neural correlate, of the change of the
16 But not their subatomic constituents (electrons, nucleons, quarks). This does not mean

that under very special laboratory conditions entire molecules cannot find themselves as a
whole in a superposed quantum statediffraction experiments with C 60 molecules have been
carried out successfully (Hackermuller et al. 2004).

14
macroscopic state of the apparatus as a result of the measurement (knowledge
of the value of the binary observable).
Following the usual von Neumann protocol, let us call |Mi the initial state of
the apparatus and |M0 i,|M1 i the two possible alternative states of the apparatus
after the measurement. The instrument is deliberately built in such a way that
when the qubit to be measured is in basis state |0i before its interaction with the
apparatus, the final independent state of the instrument after the measurement
will be |M0 i, and if the state of the qubit is |1i the instrument will end up in state
|M1 i. In either case, the state of the qubit remains unchanged (we are assuming
this to be a non-destructive measurement). Therefore, for the composite state
qubitapparatus we will have the following evolution in time, as determined by
the Schrodinger equation: |0i |Mi |0i |M0 i or |1i |Mi |1i |M1 i.
2 2
If the qubit is now in a superposed state |0i + |1i (with || +|| =1),
since the Schrodinger equation is first-order in time we will obtain an entan-
gled state and the state of the composite system qubitapparatus will remain
a linear superposition, as long as it is kept isolated from all other interactions:
(|0i+|1i)|Mi |0i|M0 i+|1i|M1 i. However, it is an experimental fact that
one has never observed a macroscopic system with such peculiar properties as
superposition (the essence of the Schrodinger cat paradox): the end state of the
composite system will always be either |0i|M0 i or |1i|M1 i (decoherence) with
2 2
|| and || the probabilities to obtain either result, respectively, in a large
set of measurements under strictly identical conditions from preparation to end
result (Born rule). In each process, the instrument comes out in the macro-
scopic state |M0 i or |M1 i, and the original qubit emerges in the corresponding
eigenstate (state reduction).
According to our definition of pragmatic information, decoherence and state
reduction thus express the fundamental fact that no information can be extracted
experimentally on the superposed state of a single qubit. A direct consequence
of this is the fact, already mentioned before, that the state of one given qubit
cannot be copied. Indeed, if we could make N ( ) copies of a single qubit
in a superposed state, a correspondence could indeed be established between
the original pair , and some macroscopic feature linked to the statistical
outcome of measurements on the N copies (in the case of a qubit, for instance
the average values of some appropriate observables). This would be tantamount
to extracting pragmatic information from the original single qubit. What is
possible, though, is to repeat exactly N times the preparation process to obtain
N separate qubits in the same superposed state (like retyping a text on blank
sheets repeatedly, instead of Xeroxing the original N times). Each qubit of this
set can then be subjected to a measurement, and the parameters , extracted
2
from the collection of results (this is, precisely, how the probabilities || and
2
|| are obtained experimentally17 ).
The preparation of a qubit in a given superposed state requires the inter-
17 Remember that since and are two normalized complex numbers, in order to determine,

say, their relative amplitudes and phase statistically, it is necessary to obtain two sets of
measurement data (N /2 measurements in each set).

15
vention of complex macroscopic devices and three steps of action on a set of
qubits: 1) measurement of an appropriate observable (which leaves each qubit
in an eigenstate of that observable); 2) selection (filtering) of the qubits that
are in the desired eigenstate (a classical process); 3) unitary transformation (a
rotation in Hilbert space) to place the selected qubit in the desired superposed
state. In this procedure a preparer has converted a certain macroscopic pattern
(embedded in the physical configurations of the preparation process) into the
values of two complex parameters of a quantum system (e.g., the normalized
and coefficients in the mathematical description of the qubit). According
to our definition, doesnt such correspondence represent genuine pragmatic in-
formation (correspondence pattern change)? No, because it would not be
univocal: given a single qubit in an unknown state, an observer could never
reconstruct through measurement the original macroscopic configuration used,
or steps taken, in the preparation process. The only way to do so is to remain in
the macroscopic domain and ask the preparer (classical information from brain
to brain!). Once known via such a macroscopic route, it will in principle be
possible to verify (but not to determine from scratch) the information about
and .

4.3 Entangled qubits and space-time

Take two qubits A and B that are maximally entangled in the antisymmetric
Bell state = 1/ 2(|0iA |1iB |1iA |0iB ) at time t0 . We may imagine qubit
B now being taken far away. If nothing else is done to either, we can bring B
back, and with some suitable experiment (e.g., interference) demonstrate that
the total state of the system had remained entangled all the time. If, instead, at
time tA > t0 a measurement is made on qubit A leaving the composite system
reduced to either state |0iA |1iB or |1iA |0iB , qubit B will appear in either basis
state |1iB or |0iB , respectively, if measured. The puzzling thing is that it does
not matter when that measurement on B is madeeven if made before tA . Of
course, we cannot predict which of the two alternatives will result; all we can
affirm is that the measurement results on each qubit will appear to be correlated,
no matter the mutual spatial distance and the temporal order in which they were
made18
The result of all this is that it appears as if the reduction of the quantum
state of an entangled system triggered by the measurement of one of its com-
ponents is non-local in space and time. Yet as stated before, correlation does
not mean causation in the quantum domain: nothing strange happens at the
macroscopic level: the state reduction cannot be used to transmit any real infor-
mation from A to B. In terms of our definition of pragmatic information, there
is no spooky action-at-a-distance: an experimenter manipulating A has no
control whatsoever over which macroscopic change shall occur in the apparatus
18 One could argue that in the case of a measurement on B at an earlier time t < t , it was
B A
this measurement that caused the reduction of the qubits composite statebut the concepts
of earlier and later between distant events are not relativistically invariant properties. See
also next paragraph.

16
at B, and vice versa. The spookiness only appears when, in the mental image
of a pair of spatially separated entangled qubits, we force our (macroscopic)
concept of information into the quantum domain of the composite system and
think of the act of measurement of one of the qubits as causing the particular
outcome of the measurement on the other.
Yet another insight can be gleaned from the re-examination of the so-called
quantum teleportation of a qubit (e.g., Bouwmeester et al., 1997). Let us re-
member the basic procedure:
an entangled pair of qubits in the antisymmetric
Bell state = 1/ 2(|0iA |1iB |1iA |0iB ) is produced at time t0 and its com-
ponents are taken far away from each other. At time tA > t0 an unknown
qubit in superposed state |0iC + |1iC is brought in and put in interaction
with
qubit A. The total, composite, state of the three-qubit system is now
1/ 2(|0iC + |1iC )(|0iA |1iB |1iA |0iB ), which can be shown algebraically to
be equal to a linear superposition of four Bell states in the A-C subspace, with
coefficients that are specific unitary transforms of the type (|0iB |1iB ),
(+|1iB |0iB ), and so on. Therefore, if a measurement is made on the pair
A-C of any observable whose eigenstates are the four Bell states, the state of
the entire system will collapse into just one of the four terms, with the qubit at
B left in a superposed state with coefficients given by the parameters , of the
now vanished unknown qubit C. If the observer at point A informs B (a classi-
cal, macroscopic transfer of pragmatic information) which basis Bell state has
resulted in the measurementonly two bits are needed to label each possible
basis stateobserver B can apply the appropriate inverse unitary transforma-
tion to his qubit, and thus be in possession of the teleported qubit C (defined
by the unknown coefficients and ).
The puzzling aspect of this procedure is that it looks as if the infinite amount
of information on two real numbers (those defining the normalized pair of com-
plex numbers and ) was transported from A to B by means of only two classi-
cal bits. The answer is that, again, according to our definition, , do not repre-
sent pragmatic information on the state of any given qubit. They are quantita-
tive parameters in the mathematical framework developed to describe quantum
systems and their interactions, but they cannot be determined physically out
there for a given qubit. Related to this, there is no way to verify the teleporta-
tion of a single qubit; the only way verification could be accomplished is through
a statistical process, repeating the whole procedure N times, from the identical
preparation of each one of the three qubits A, B, C to the actual measurement
of the teleported qubit. If we determine the frequencies of occurrence N0 and
N1 (= N N0 ) of the |0iB and |1iB states, and express the teleported qubit
state in its polar (Bloch sphere) form |i = cos(/2)|0i + exp(i) sin(/2)|1i,
it can be shown that the number of statistically significant figures (in base 2)
of and to be obtained is equal to the total number 2N of bits transmitted
classically (i.e., macroscopically) from A to B, so that from the statistical point
of view, there is no puzzle at all! Moreover, this shows that there is no way of
teleporting pragmatic information and, as a consequence, macroscopic objects!
The preceding discussion says something about how we tend to think intu-
itively of time and space at the quantum level. A point in the 4-dimensional

17
continuum of space-time is a mathematical abstraction, useful in the description
of objects and events in the Universe. But position and time intervals of objects
must be determined by measurements, i.e., with a macroscopic instrument, in
which a macroscopic change is registered (e.g., a change in the bath of light in
a position measurement; a change in the configuration of a clock or the angular
position of a star). Like information, time is a macroscopic concept (even an
atomic clock must have classical components to serve as a timepiece). We can
assign time marks to a quantum system only when it interacts locally with (or
is prepared by) a macroscopic system. In the case of a wave function (x, t),
the time variable refers to the time, measured by a macroscopic clock external
to the quantum system, at which ||2 is the probability density of actually ob-
serving a quantum system at the position x in configuration space, which is also
based on a measurement with a macroscopic instrument.
Non-locality in space and time really means that for a composite quantum
system, the concepts of distance and time interval between different superposed
or entangled components are undefined as long as they remain unobserved, i.e.,
free of interactions with macroscopic systems19 (for a postulate on atemporal
evolution, see Steane, 2007). Because of this, it may be unproductive trying
to find a modified form of the Schrodinger equation, or any other formalism,
to describe quantitatively what happens inside a quantum system during the
process of state reduction.20 But, finally, what about the decay process of an
unstable particle (or nucleus, for that matter), which run on the proper time of
the particle, as demonstrated long ago with the decay times of cosmic ray mu-
mesons when observed from a reference frame fixed to Earth? As we shall briefly
mention, decay processes may be linked to decoherence; if that is indeed the case,
they would be controlled by interactions with the macroscopic environment as
experienced by the quantum system (reduction to decay products).

4.4 The process of decoherence

Returning to decoherence, let me briefly address the still contentious question
of what, if any, physical processes are responsible for the transition from the
quantum end of a measurement apparatus to its classical, observable one (for
details, see, e.g., Schlosshauer 2008). For this purpose, let us consider a super-
simplified model of a measuring apparatus that consists entirely of mutually
interacting qubitslots of them, perhaps 1022 or 1023 with one of which the
external qubit to be measured enters into unitary non-destructive interaction at
time t0 . In our model, the apparatus qubits represent a complex web in some
initial or ready state, designed in such a way that, as the local unitary inter-
action processes multiply and propagate, only two distinguishable macroscopic
end states M0 and M1 can be attained, realized as two macroscopic spatially
19 Already in 1927 Heisenberg declared: A particle trajectory is created only by the act of

observing it!
20 For instance, a theoretical or experimental derivation of average trajectories in a double-

slit experiment (e.g., Kocsis et al. 2011) provides the geometric visualization of something on
which, for an individual particle, pragmatic information could never be obtained!

18
or temporally different forms or patterns (the so called pointer states, repre-
sented in mutually orthogonal Hilbert subspaces of enormous dimensions). The
key physical property of this construction is that for a qubit in a basis state
the instruments final macroscopic configuration will depend on the actual basis
state of the measured qubit (see discussion in subsection 4.2).
If the qubit to be measured is now in a superposed state, the first physical
interaction at time t0 would create an entangled state of the system qubitfirst
apparatus quantum element which through further unitary inter-component in-
teractions would then expand to the entire composite system qubitapparatus
in a cascade of interactions and further entanglements throughout the appara-
tus. Accordingly, the classical end of the apparatus also should end up in a
superposed state (the Schrodinger cat!) and since in principle the interactions
are unitary, the whole process would be reversible. Obviously, somewhere in
the cascade of interactions there must be an irreversible breakdown of the en-
tanglement between the original qubit and the instrument, both of which will
emerge from the process in separate but correlated states, either |0i and M0 ,
or |1i and M1 , respectively. Extractable pragmatic information appears only at
the classical end of this cascade. All this means that if exactly the same kind
of measurement is repeated immediately on the qubit (assuming that it was
not destroyed in the measurement process), one will obtain a result that with
certainty is identical to that of the original measurement.
Measurement processes and their apparatuses are artificeshuman planned
and designed for a specific purpose. However, note that the preceding discus-
sion on decoherence can be applied to the case in which we replace the artificial
measurement apparatus with the natural environment per se. Just replace the
word apparatus with the environment with which a given quantum system
willy-nilly interacts and gets entangled. As long as this entanglement persists,
the given quantum system will have lost its original separate state, and only
the composite quantum systemenvironment will have a defined state, how-
ever complicated and delocalized. Now, if the interaction with the environment
leads to a macroscopic change somewhere (potentially verifiable through clas-
sical information-extraction by an observer, but independently of whether such
verification actually is made), it will mean that decoherence has taken place and
that the state vector of the original quantum system will have been reduced to
one of its original eigenstates pertinent to the particular interaction process.
This is usually described as entanglement with the environment carrying
away information on a quantum system, or information about the systems
state becoming encoded in the environment. However, I would like to caution
about the use of the term information in this particular context: there is no loss
of pragmatic information in natural decoherence, because there wasnt any there
in the first place! A much less subjective way is to say: A quantum system con-
tinuously and subtly interacts with its environment and gets entangled with it; if
decoherence occurs, a macroscopic change in the state of the environment will
appear somewhere (information about which could eventually be extracted by an
observer), and the state of the quantum system will appear reduced to some spe-
cific basis state in correspondence with the environmental change in question.

19
Since this basis state is in principle knowable, the decohered qubit now belongs
to the classical domain. Measurement instruments are environmental devices
specifically designed to precipitate decoherence and steer it to into certain sets
of possible final macroscopic states (preferred states, if no alternative sets exist).
An ensemble of identically prepared quantum systems (e.g., a chunk of a re-
cently separated, chemically pure radioisotope) thus turns probabilistic because
it is unavoidably submerged in a gravitating, fluctuating, thermodynamic
macro-world, and will decay into a mixture of quantum entities in eigenstates
21
(e.g., with an -particle either still inside or already outside a nucleus).
On the other hand, the laboratory measurement of a quantum system may be
viewed as a case in which the environment was deliberately altered by inter-
posing a human-made apparatus, which then altered in a not-so-subtle way
the time evolution of the system (we should really say: the time evolution of
potential macroscopic effects of the system) by greatly increasing the chance of
decoherence. In summary, what we have called the cascade of entanglements in
a quantum measurement also involves a stochastic ensemble of outer environ-
mental components with which the instruments components are in subtle but
unavoidable interaction.
A collateral consequence of natural decoherence is that any peculiar quan-
tum property like superposition will have little chance of spreading over a ma-
jor part of a macroscopic object, which indeed will behave classically whenever
observedthere always seems to be a natural limit to the complexity of a quan-
tum system in a pure superposed state beyond which it will decohere. In other
words, the classical macroscopic domain, in which life systems operate and in-
formation can be defined objectively, consists of objects whose constituents have
decohered into eigenstates (mainly, of their Hamiltonians). Quantum behavior
of a macroscopic system is not forbidden (a Schrodinger cat could be in a super-
posed state of dead and alive at the same time!), but its probability and duration
would be ridiculously small. This also explains the fact that, as mentioned be-
fore, many artificial quantum systems are very unstable in a superposed state,
and thus very difficult to handle in the laboratorya fact that represents one
of the biggest challenges to quantum computing.
Finally, we should view all dynamics equations in quantum mechanics like
the Schrodinger equation as the tools for providing information on potential
macroscopic effects of a quantum system on the environment (or a measurement
apparatus) under given circumstances, rather than describing the evolution of
the quantum system per se. The amazing aspect of quantum mechanics is
not its puzzling paradoxes, but the fact that a mathematical framework could
be developed that can be used successfully to determine statistical, objective
probabilities for observable macroscopic outcomes of interaction processes both,
in artificial experiments and in the natural environment.
21 With the decay times mainly depending on the original wave function of the individual

nuclei, but also slightly perturbed by subtle but unavoidable interactions of the latter with
the environment (leading to fluctuations in the exponential decay of the ensemble.)

20
5 Concluding Remarks: Quantum Pedagogy
From the previous discussion it is advisable to refrain from using the classi-
cal concept of pragmatic information indiscriminately to represent mentally a
quantum systembe it by thinking about it, mathematically representing it or
manipulating it in Gedankenexperiments. Yet quite commonly we do, especially
when we teachbut then, as mentioned above, we should not be surprised that
by forcing the concept of information into the quantum domain, mental im-
ages are triggered of weird behavior that is contradictory to our every-day
macroscopic experience.
To me, the problem of the interpretation of quantum mechanics is not just
one of a philosophical nature but one of eminently pedagogical nature. For
instance, how should one answer correctly the often-asked question: Why is it
not possible, even in principle, to extract information on the actual state of a
single qubit? Because by the definition of information, to make that possible
there would have to exist some physical paradigm by means of which a change is
produced somewhere in the macroscopic classical domain that is in one-to-one
correspondence with the qubits parameters immediately prior to that process.
Only for eigenstates (basis states) can this happendecoherence prevents the
formation of any macroscopic trace of superposed states. In the case of an
initially superposed state, the end state of the qubit will always appear corre-
lated with the end state of the macroscopic system, i.e., will emerge reduced
to a correlated or preferred basis state. In somewhat trivial summary terms,
quantum mechanics can only provide real information on natural or deliber-
ate macroscopic imprints left by a given quantum system that has undergone a
given preparation, eventually interact unitarily (reversibly) with other quantum
systems forming a composite quantum system, which as a single whole interacts
irreversibly with the surrounding macroscopic world.
So what are the coefficients in a qubit state like |0i + |1i or |i =
cos(/2)|0i + exp(i) sin(/2)|1i? They are parameters in a model represen-
tation of the system in complex Hilbert space, which within an appropriate
mathematical framework enables us to make quantitative, albeit only proba-
bilistic, predictions about the systems possible macroscopic imprints on the
classical domain. We may call and or and information, and we do,
based on the fact that we can prepare a quantum system in a chosen superposed
statethe common usage of the terms quantum bit and quantum informa-
tion testifies to this. Yet for a single qubit we cannot retrieve, copy or verify
the numbers involved, which means we cannot establish a univocal correlation
between the state of the qubit and any macroscopic feature. In other words,
those parameters are not pragmatic information (the only exception is when
the qubit is in one of its basis states). This is why when we do call , , and
information, we are always obliged to point out its hidden nature! And
in teaching, we always would have to mumble something about super-luminal
speed of information, teleportation of real things, a particle being in different
positions at the same time, etc. to satisfy our (classical world) imagination.
In Richard Feynmans words, we always would have to emphasize that ...the

21
[quantum] paradox is only a conflict between reality and your feeling of what
reality ought to be. The whole framework of quantum information theory and
computing is based on the consistency of this kind of classical correspondence:
the relation of a given initial set of qubits in prepared eigenstates (a classical
input pattern) correlated through intermediate unitary quantum interactions in
an appropriately shielded quantum computing device with another final set of
qubits in basis states (a classical output pattern). This input-output correla-
tion is what really should called quantum information, a genuine category of
pragmatic information.
Obviously, during the time interval between input and output, any extra-
neous non-unitary intervention, whether artificial (a measurement) or natural
(decoherence), will change or destroy the macroscopic input-output correlation.
Indeed, in this interim interval, the proverbial mandate of dont ask, dont tell
applies (Roederer 2005)not because we dont know how to extract relevant
information to answer our questions, but because pragmatic information per se
does not operate in the quantum domain.
Let me end by stating a personal opinion as a former physics teacher. When
it comes to evaluating, or to teaching about, the realities out there, it is the
physicists who should get real and recognize the fact that the World, both
physical and biological, does not operate on the basis of what happens in Mach-
Zehnder interferometers, Stern-Gerlach experiments, two-slit diffraction labora-
tory setups, qubit teleportation and all these marvellous experiments designed
and performed by humans. In reality, all such experiments, while providing
answers to the inborn human inquiry about how our environment works, are
but artificial intrusions poking into a Universe that does not care about linear
algebra, Hamiltonians, and about information per se. These experiments and
the ensuing human understanding were made possible only because of the emer-
gence, at least on Planet Earth, of interactions based not on force fields alone,
but on the evolution of ultra-complex macroscopic mechanisms responding ex-
clusively to simple geometric patterns in space and time.

6 References
Balter, M. 2010. Did working memory spark creative culture? Science 328:
160163.
Bickerton, D. 1995. Language and Human Behavior. Seattle, WA: University
of Washington Press.
Bouwmeester, D.; Pan J. W.; Mattle, K.; Eibl, M.; Weinfurter, H.; and Zeilinger,
A. 1997. Experimental quantum teleportation. Nature 390(6660): 575-579.
Fuchs, C. A.; Mermin, N. D.; and Schack, R. 2013. An introduction to Qbism
with an application to the locality of quantum mechanics. arXiv: 1311.5253v1
(quant-ph).
Han, C. J.; OTuathaigh, C. M.; van Trigt, L.; Quinn, J. J.; Fanselau, M. S.;
Mongeau, R.; Koch, C.; and Anderson, D. J. 2003. Trace but not delay fear
conditioning requires attention and the anterior cingulated cortex. Proc. Natl.

22
Acad. Sci. USA 100(22): 13087-13092.
Hackermuller, L.; Hornberger, K., Brezger, B.; Zeilinger, A.; and Arndt, M.
2004. Decoherence of matter waves by thermal emission of radiation. Nature
427 : 711-714.
Kuppers, B.-O. 1990. Information and the Origin of Life. Cambridge, Mass.:
The MIT Press.
Kocsis, S.; Braverman, B.; Ravets, S.; Stevens, M. J.; Mirin, R. P.; Shalm, L.
K.; and Steinberg, A. M. 2011. Observing the average trajectories of single
photons in a two-slit interferometer. Science 332 : 1170-1173.
Mermin, N. D. 2007. Quantum Computer Science. Cambrisge University Press,
New York.
Roederer, J. G. 1978. On the relationship between human brain functions and
the foundations of physics. Found. of Phys. 8 : 423-438.
Roederer, J.G. 2003. Information and its role in nature. Entropy 5: 1-31.
Roederer, J.G. 2004. When and where did information first appear in the Uni-
verse? In New Avenues in Bioinformatics J. Seckbach et al., eds. Kluwer Acad.
Publ., Dordrecht.
Roederer, J. G. 2005. Information and its Role in Nature. Berlin, Heidelberg,
New York: Springer-Verlag.
Roederer, J. G. 2012. Toward an information-based interpretation of quantum
mechanics and the quantum-classical transition. arXiv: 1108.0999v2 (quant-
ph).
Schlosshauer, M. 2008. Decoherence and the Quantum-to-Classical Transition.
Berlin, Heidelberg: Springer-Verlag.
Shannon, C. E., Weaver W. W. 1949. The Mathematical Theory of Communi-
cation. University of Illinois Press.
Steane, A. M. 2007. Context, spacetime loops and the interpretation of quantum
mechanics. J. Phys. A: Math. and Theor. 40: 3223-3243.
Tononi, G. and Koch, C. 2008. The neural correlates of consciousness: an
update. Annals of the New York Academy of Science 1124 : 239-261.
Walker, S. I., Kim, H. and Davies, P. C. W. 2016. The informational architecture
of the cell. In press Phil. Transactions.
Wheeler, J. A. 1989. Information, physics, quantum: The search for links. Proc.
3rd Int. Symp. Foundations of Quantum Mechanics, Tokyo, 1989, pp.354-368.

23