1 Introduction to Nonequilibrium Statistical Physics and Its Foundations
3
Let us discuss how come objectivity is possible, despite inevitable subjective interference in all possible measurements. Indeed, as measurement tools and protocols
are chosen by someone, who may want to investigate a given phenomenon and not
others, the objectivity of data may look problematic.
The most familiar measurement tools are afforded by our senses: amazingly complex structures, more accurate and dependable than many artificial sophisticated
tools. Nevertheless, our sense are imperfect and may badly lead us into error, like
hallucinations or misjudgements. Even when such negative situations do not happen,
our senses may still induce different people to draw contradictory conclusions on the
very same phenomenon. For instance, natt ¯
o rotten beans, considered a delicacy in
Japan, result disgusting to many westerners.
Measurement tools are extensions of our senses, meant to reduce this degree
of subjectivity, by increasing accuracy and resolution. Presently, the duration of a
second can be split in 10
16 parts, the length of a forearm in 10
10 parts, the weight
of a pollen grain in 10
18 parts. This may look sufficiently accurate to eliminate any
ambiguity and source of subjectivity in physical measurements. Nevertheless, given
an object of investigation, the observer shall decide which aspect of the object should
be subjected to measurement, which tools shall be used, which protocols shall be
adopted. Moreover, while it is obvious that different quantities, such as time and
mass, require different tools to be measured, it is interesting to observe that a single
quantity may require different tools, depending on the scale of observation: one
cannot weigh a star with a scale built on Earth, while one such scale can weigh
diamonds.
Even agreeing on the quantities to be measured and on the tools to be used, people
can still get to totally different conclusions. Let us consider a cloud and use our eyes.
A common experience of travellers is that the cloud looks like a white object with a
definite shape and extension in the blue sky, but flying closer and closer, the cloud
boundaries gradually vanish, and after a while there is only grey: no white, no blue.
Which picture is more real? The white shape in the blue sky, or the grey that seems
to have neither a beginning nor an end? Which of the measurement is more accurate?
The one performed before take off, or that performed right inside the cloud?
One could conclude that there are too many points of view on a very same object,
hence plenty of room for subjectivity rather than objective science. This would seem
to undermine the foundations of Galilean science, which has led progress for centuries, even before the recent attacks from big data and machine learning got to the
stage. But this is not the case. As Chinese culture teaches, the truth needs the apparently contrasting views to be harmonized. There is room for subjectivity in deciding
what is interesting to measure: we may choose to classify the colors of objects, rather
than their masses. There is room for objectivity in using the radiation wavelengths
for colors, and pounds for masses.
Indeed, remember Rosetta and its adventure in space
1 : it took eleven years to
objectively reach the 67P-Churyumov-Gerasimenko comet, and to objectively land
the probe Philae on it. It was all based on an objective understanding of the law of
1 See e.g. https://www.nasa.gov/rosetta.
3
Let us discuss how come objectivity is possible, despite inevitable subjective interference in all possible measurements. Indeed, as measurement tools and protocols
are chosen by someone, who may want to investigate a given phenomenon and not
others, the objectivity of data may look problematic.
The most familiar measurement tools are afforded by our senses: amazingly complex structures, more accurate and dependable than many artificial sophisticated
tools. Nevertheless, our sense are imperfect and may badly lead us into error, like
hallucinations or misjudgements. Even when such negative situations do not happen,
our senses may still induce different people to draw contradictory conclusions on the
very same phenomenon. For instance, natt ¯
o rotten beans, considered a delicacy in
Japan, result disgusting to many westerners.
Measurement tools are extensions of our senses, meant to reduce this degree
of subjectivity, by increasing accuracy and resolution. Presently, the duration of a
second can be split in 10
16 parts, the length of a forearm in 10
10 parts, the weight
of a pollen grain in 10
18 parts. This may look sufficiently accurate to eliminate any
ambiguity and source of subjectivity in physical measurements. Nevertheless, given
an object of investigation, the observer shall decide which aspect of the object should
be subjected to measurement, which tools shall be used, which protocols shall be
adopted. Moreover, while it is obvious that different quantities, such as time and
mass, require different tools to be measured, it is interesting to observe that a single
quantity may require different tools, depending on the scale of observation: one
cannot weigh a star with a scale built on Earth, while one such scale can weigh
diamonds.
Even agreeing on the quantities to be measured and on the tools to be used, people
can still get to totally different conclusions. Let us consider a cloud and use our eyes.
A common experience of travellers is that the cloud looks like a white object with a
definite shape and extension in the blue sky, but flying closer and closer, the cloud
boundaries gradually vanish, and after a while there is only grey: no white, no blue.
Which picture is more real? The white shape in the blue sky, or the grey that seems
to have neither a beginning nor an end? Which of the measurement is more accurate?
The one performed before take off, or that performed right inside the cloud?
One could conclude that there are too many points of view on a very same object,
hence plenty of room for subjectivity rather than objective science. This would seem
to undermine the foundations of Galilean science, which has led progress for centuries, even before the recent attacks from big data and machine learning got to the
stage. But this is not the case. As Chinese culture teaches, the truth needs the apparently contrasting views to be harmonized. There is room for subjectivity in deciding
what is interesting to measure: we may choose to classify the colors of objects, rather
than their masses. There is room for objectivity in using the radiation wavelengths
for colors, and pounds for masses.
Indeed, remember Rosetta and its adventure in space
1 : it took eleven years to
objectively reach the 67P-Churyumov-Gerasimenko comet, and to objectively land
the probe Philae on it. It was all based on an objective understanding of the law of
1 See e.g. https://www.nasa.gov/rosetta.