76
The position that the purpose of measurement is to discover objectively existing
values of quantities extends well beyond the scientific community. For example,
such a position helped motivate the introduction of the metric system during the
French revolution, as described by Witold Kula (1986: p. 123):
The meter, in ‘dehumanizing’ measures, in rendering them independent of man and ‘objective’ in their interrelationship with man as well as morally neutral, has also transformed an
instrument of ‘man’s inhumanity to man’ into a means of understanding and cooperation
for mankind.
Indeed, measurement results had been used in the past to give a social justification
for decisions that otherwise could seem capricious; in the years following the first
period of the French revolution, many asked “what is the use to us of the abolition
of the feudal system, if the seigneurs remain at liberty arbitrarily to increase or
decrease their measures?” (p. 234). This refers to misuses of conventionality in the
pre-Napoleonic choice of measurement units, at the same time implying that it must
be possible to have just and fair measurement systems, and that a step towards this
goal is to remove individuals’ ability to arbitrarily “increase or decrease […]
measures”.
Yet even though the interpretation of measurement as aimed at the discovery of
independently existing values is attractive for its simplicity, it falls short precisely in
that it equates measurement with a transmission process. Consider the nature of
communication and transmission, as they are technically understood: communication is a mapping of one informational entity to another informational entity, performed through transmission, which is a mapping of one empirical entity to another
empirical entity, e.g., electromagnetic waves (Shannon & Weaver, 1949). That is,
the input to a communication system is an informational entity, explicitly provided
by an agent who (or which) operates on purpose by encoding the informational
entity into an empirical property of the transmission system, which is then transferred to the receiver and finally decoded into an informational entity. For example,
a text may be encoded into an appropriately modulated amplitude of an electric
signal, but also, trivially, into a sheet of paper with patterns of ink on it; the electric
signal or the sheet of paper is transferred, and then decoded into a text. The informational entity transmitted along the transmission channel via the encoded empirical property is, at least in principle, perfectly knowable, given that it was purposely
generated or selected by an agent. No such agent exists in the case of measurement,
which in turn may be modeled as a mapping from an empirical entity to an informational entity: in a measurement process, the values of properties are the output, not
the input. This essential difference between communication and measurement is
depicted in Fig. 4.1.
By neglecting this difference, naïve realism misses some key features of measurement and is unable to account for the role of models in the definition of the
measurand, in the development of the measurement procedure, and in the interpretation of the results of its application (e.g., Frigerio, Giordani, & Mari, 2010; McGrane
& Maul, 2020; Teller, 2013).
4 Philosophical perspectives on measurement
The position that the purpose of measurement is to discover objectively existing
values of quantities extends well beyond the scientific community. For example,
such a position helped motivate the introduction of the metric system during the
French revolution, as described by Witold Kula (1986: p. 123):
The meter, in ‘dehumanizing’ measures, in rendering them independent of man and ‘objective’ in their interrelationship with man as well as morally neutral, has also transformed an
instrument of ‘man’s inhumanity to man’ into a means of understanding and cooperation
for mankind.
Indeed, measurement results had been used in the past to give a social justification
for decisions that otherwise could seem capricious; in the years following the first
period of the French revolution, many asked “what is the use to us of the abolition
of the feudal system, if the seigneurs remain at liberty arbitrarily to increase or
decrease their measures?” (p. 234). This refers to misuses of conventionality in the
pre-Napoleonic choice of measurement units, at the same time implying that it must
be possible to have just and fair measurement systems, and that a step towards this
goal is to remove individuals’ ability to arbitrarily “increase or decrease […]
measures”.
Yet even though the interpretation of measurement as aimed at the discovery of
independently existing values is attractive for its simplicity, it falls short precisely in
that it equates measurement with a transmission process. Consider the nature of
communication and transmission, as they are technically understood: communication is a mapping of one informational entity to another informational entity, performed through transmission, which is a mapping of one empirical entity to another
empirical entity, e.g., electromagnetic waves (Shannon & Weaver, 1949). That is,
the input to a communication system is an informational entity, explicitly provided
by an agent who (or which) operates on purpose by encoding the informational
entity into an empirical property of the transmission system, which is then transferred to the receiver and finally decoded into an informational entity. For example,
a text may be encoded into an appropriately modulated amplitude of an electric
signal, but also, trivially, into a sheet of paper with patterns of ink on it; the electric
signal or the sheet of paper is transferred, and then decoded into a text. The informational entity transmitted along the transmission channel via the encoded empirical property is, at least in principle, perfectly knowable, given that it was purposely
generated or selected by an agent. No such agent exists in the case of measurement,
which in turn may be modeled as a mapping from an empirical entity to an informational entity: in a measurement process, the values of properties are the output, not
the input. This essential difference between communication and measurement is
depicted in Fig. 4.1.
By neglecting this difference, naïve realism misses some key features of measurement and is unable to account for the role of models in the definition of the
measurand, in the development of the measurement procedure, and in the interpretation of the results of its application (e.g., Frigerio, Giordani, & Mari, 2010; McGrane
& Maul, 2020; Teller, 2013).
4 Philosophical perspectives on measurement