255
relevant stakeholders.
11
What was mentioned above about science can be repeated
about measurement, then: it is no contradiction to say that a given process is both a
measurement and produces results that are of low quality (i.e., is a bad measurement) but it would be a contradiction to say that a process is both a measurement
and cannot be justified, even if its results are (perhaps accidentally) accurate. Of
course, in many cases, particularly of nonscientific measurements, measurement
systems remain black boxes and no justifications of their results are reported: what
is required is that a justification can be provided, whenever required.
On this basis, we are finally ready to propose a characterization of that includes both necessary and sufficient conditions:
measurement is an empirical and informational process that is designed on purpose, whose
input is an empirical property of an object, and that produces explicitly justifiable information in the form of values of that property
In this way, the expectations of good measurement processes are continuous with
the expectations of other sources of knowledge, in Plato’s sense of knowledge as
justified true belief. This is the case regardless of the subject matter and the details
of the procedures involved in measurement.
8.3.5 Consequences for the theory and the practice
of measurement
The requirement that measurement results be explicitly justifiable helps explain
why measurement cannot be adequately characterized solely using a black box
model: if a given attribution of value(s) to a property is claimed to be a measurement
(instead of, e.g., once again, a guess), it must be possible to explain how it was performed, and this requires opening the box and identifying the features of the process
that secure the quality of the results. Given the diverse contexts in which measurements are applied, it is not surprising that what is found inside the (metaphorical,
but sometimes also actual) box is also diverse.
We claim that there is a commonality in this diversity, however, and this commonality is structural: no matter how complex is the measurement system, (i) it is
based on one or more direct measurements, as discussed in Sect. 7.2, and (ii) the
structure of a direct measurement is based on the Hexagon Framework, introduced
in Sect. 7.3. The fundamental point here is that it is the structure itself of the process
11 A more specific condition is then that measurement results need to be reproducible by all relevant
social stakeholders. Even neglecting the practical constraints related to the fact that setting up a
measurement system may have costs which are not affordable for all interested parties, we must
acknowledge that some measurements are not repeatable, for example when they alter the state of
the object under measurement in an irreversible way (see, e.g., destructive testing, www.electropedia.org/iev/iev.nsf/display?openform&ievref=151-16-29). Hence reproducibility cannot be taken
as a characterizing condition.
8.3 Can there be one meaning of “measurement” across the sciences?
relevant stakeholders.
11
What was mentioned above about science can be repeated
about measurement, then: it is no contradiction to say that a given process is both a
measurement and produces results that are of low quality (i.e., is a bad measurement) but it would be a contradiction to say that a process is both a measurement
and cannot be justified, even if its results are (perhaps accidentally) accurate. Of
course, in many cases, particularly of nonscientific measurements, measurement
systems remain black boxes and no justifications of their results are reported: what
is required is that a justification can be provided, whenever required.
On this basis, we are finally ready to propose a characterization of
measurement is an empirical and informational process that is designed on purpose, whose
input is an empirical property of an object, and that produces explicitly justifiable information in the form of values of that property
In this way, the expectations of good measurement processes are continuous with
the expectations of other sources of knowledge, in Plato’s sense of knowledge as
justified true belief. This is the case regardless of the subject matter and the details
of the procedures involved in measurement.
8.3.5 Consequences for the theory and the practice
of measurement
The requirement that measurement results be explicitly justifiable helps explain
why measurement cannot be adequately characterized solely using a black box
model: if a given attribution of value(s) to a property is claimed to be a measurement
(instead of, e.g., once again, a guess), it must be possible to explain how it was performed, and this requires opening the box and identifying the features of the process
that secure the quality of the results. Given the diverse contexts in which measurements are applied, it is not surprising that what is found inside the (metaphorical,
but sometimes also actual) box is also diverse.
We claim that there is a commonality in this diversity, however, and this commonality is structural: no matter how complex is the measurement system, (i) it is
based on one or more direct measurements, as discussed in Sect. 7.2, and (ii) the
structure of a direct measurement is based on the Hexagon Framework, introduced
in Sect. 7.3. The fundamental point here is that it is the structure itself of the process
11 A more specific condition is then that measurement results need to be reproducible by all relevant
social stakeholders. Even neglecting the practical constraints related to the fact that setting up a
measurement system may have costs which are not affordable for all interested parties, we must
acknowledge that some measurements are not repeatable, for example when they alter the state of
the object under measurement in an irreversible way (see, e.g., destructive testing, www.electropedia.org/iev/iev.nsf/display?openform&ievref=151-16-29). Hence reproducibility cannot be taken
as a characterizing condition.
8.3 Can there be one meaning of “measurement” across the sciences?