The mode of defining in contemporary physical science

In commenting on the absolute velocity of the speed of light, which is at the basis of Einstein’s relativity, Eddington notes that we need to understand the claim “nothing can go faster than the speed of light” in a particular way. If we build a laser cannon and shine it to Neptune, then spin the cannon around in a circle at even a moderate speed, the tip of the light beam will be moving much faster than the speed of light relative to us. Eddington says that the theory only deals with speeds that are capable of being signals. Einstein himself makes a point that dovetails with this in his discussion of simultaneity in Relativity, when he specifies the peculiar character of physical definitions (that they have to show us how to verify something in a particular case), and then tells the reader “not to read any further until they are firmly convinced of the point.” This is to say we must define the things we study in terms of signals that can be verified and detected in a particular case. This mode of defining runs across all sciences from physics to statistics and poll gathering to chemistry to sociology to psychology. The whole nature of modern science – with all its power and limitations – rests on seeing this peculiar mode of defining.


  1. Brandon said,

    June 22, 2011 at 9:40 am

    This links up with what is the truth at the core of the more sane kind of positivisms (e.g., Duhem’s).

  2. Peter said,

    June 22, 2011 at 3:45 pm

    How would you define definition in the sense that it applies to modern sciences? (The Aristotlelian definition of definition as “speech making known what a thing is”, doesn’t exactly fit.)

  3. Brandon said,

    June 22, 2011 at 5:59 pm

    Wouldn’t the Aristotelian definition still fit under a qualification, e.g., something like, “speech making known what a thing is wholly insofar as this is detectable in signal”? Something along those lines would pin down both the sense in which a person could think it was a proper definition, and the sense in which it is actually not a proper definition.

  4. June 23, 2011 at 6:47 am

    (I’m thinking out loud here)

    Aristotle’s definitions have two components, which follow the mode of specification (defining, making distinct) of our mind: a most proximate general part (or material part. It’s tricky to pick one that is not too remote from what you are trying to define) and a part that determines it (which is composed of a number of differences and can be of indeterminate length).

    The most proximate general part of this sort of definition would be the knowledge of that which exists with what Aristotle calls “sensible matter” – which is matter so far as we can make it an object of our senses.

    We can speak of two determining parts – 1.) capable of being a signal (A signal is a sort of sign – one capable of being registered in a machine in some way). and 2.) capable of being verified in a particular instance. To get a sense of the second criteria, consider the definition of man as a rational animal. It’s perfectly true, pertains to something known with sensible matter, etc. but to know this does not give you a criterion to determine in concrete instances whether this or that thing is a man. Dictionaries are filled with these sorts of definitions. A dictionary definition of “green” won’t give you any means to verify whether, in a particular instance, some color will count as green. DeKoninck points out somewhere that the crucial word in a scientific definition is “when”, that is “if light is passed through this apparatus, green is when the wavelength numbers read between…” Clearly, the word “when” need not always be used, but you can get the idea -it’s crucial to specify some concrete process that can determine a result in a particular case. There might be some wrangling over whether this second determining factor always enters into the scientific definition, but I don’t think there is any disagreement that it needs to be the fundamental factor in attaining to any definition, if the definition is to count as scientific.

    I haven’t made up my mind whether the second factor is just an aspect of the first. “Signal”, if my definition of it is right, seems to include the second criterion.

    The idea that starts to arise from this is that science is really about what we contemporary people call information. IMO, the philosophers dealing with information theory are a hundred years ahead of the game in philosophy of science, and the guy at the head of the pack is Oliver Costa de Beauregard.

    • Peter said,

      July 8, 2011 at 4:31 am

      CDK discusses the differences in defining in his PoN Manual [draft], part 1. I would have liked to see it completed. (It wasn’t, right?)

  5. Peter said,

    June 23, 2011 at 7:20 am

    The idea of “when” is largely what I had in mind. (I have read that CDK passage, wherever it is.) It seems that in using symbols, which always reference an individual thing, or perhaps in simply always referencing an individual experiment by which the definition is verified, the determining parts of a definition as used in modern science is not universal in exactly the same way as, say, ‘rational animal’; nor does it try to be.

    In short, the fact that it is about a process done rather than directly about the universal ‘whatness’ of something strikes me as significant. “Man _is_…” as opposed to “Green _is when you_ [insert concrete action to be done]; etc..”

    And since I don’t know jack squat about information theory, I’ll just stop here….

    • John said,

      June 24, 2011 at 12:47 am

      There are two major approaches to modern physical theory. In the first a geometrical environment is postulated and the behaviour of objects predicted, this might be called the Platonic approach, and in the second, types of equation are postulated (for instance those that involve particular relations between eigenfunctions and eigenvalues) and the behaviour of objects postulated, this might be called the process, or computational approach.

      Both approaches are tested using signals because a measurement is a signal.

      Incidentally, the speed of light in modern Platonic physics is a constant in the metric of spacetime – see Wikipedia Intro to SR.

  6. Mike Flynn said,

    June 29, 2011 at 9:15 am

    In quality engineering, we call such definitions “operational definitions,” that is, the quality is defined by the operations performed in order to verify or measure it. (“Attributes” are qualities that the object possesses or not (“either/or”) while “variables” are qualities that the object possesses “more or less.” Examples being of the first: a leak in a beverage can, and of the second: a diameter of a beverage can.) W.Edwards Deming pointed out that if a quality is measured by two different series of operations, you will in general obtain two different results. For example, I have seen important differences in the measure of a dimension on a part due to the producer and the customer using two different instruments to make the measurement. Both calibrated to NIST standards.

    “There is no such thing as the speed of light,” Deming said, “only a number delivered by a specified series of operations.”

%d bloggers like this: