Monday, January 4, 2010

Are quasi-particles real?

Everyone would agree that electrons and protons are "real". But what about holes? magnons? rotons? quasi-particles in a Fermi liquid? I would say yes. It is interesting to find that this question has received some attention in the philosophical literature. One paper I found by Axel Gelfert, "Manipulative Success and the Unreal", who argues that quasi-particles counter a reality criterion due to Hacking, "If you can spray them, they exist." i.e., Gelfert claims that quasi-particles do not exist even though it is possible to manipulate them.

However, I agree with Falkenberg's counter-argument and I like her stock market analogy in the following:
quasi-particles are as real as a share value at the stock exchange. The share value is also due to a collective effect ...,namely to the the collective behavior of all investors. It is also possible to `spray' the share value in Hacking's sense, that is, to manipulate its quotation by purchase or sale for purposes of speculation. Its free fall can make an economy crash, its dramatic rise may make some markets flourish. And the crash as well as the flourishing may be local, i.e., they may only affect some local markets. But would we conclude that the share value does not exist, on the sole grounds that it is a collective effect? Obviously, share values as well as quasi-particles have another ontological status than, say, Pegasus. Pegasus does not exist in the real world but only in the tales of antique mythology. But quasi-particles exist in real crystals, as share values exist in real economies and markets. Indeed, both concepts have a well-defined operational meaning, even though their cause cannot be singled out by experiments or econometric studies.
Brigitte Falkenburg, Particle Metaphysics: a critical account of sub-atomic reality (Springer 2007), p. 245.


  1. It's interesting to think that people can assert that atoms and electrons exist while quasi-particles do not given that the evidence for any atomic or subatomic particle is in some way indirect, i.e. one measures their effects on a macroscopic instrument. In a broad sense, the properties of quasi-particles are measured in the same way, which is why I find it intriguing that some claim that quasi-particles are not "real."

    Of course, one can take my argument further and claim that all measurement is indirect, e.g. when we "see" an object we sense electrical stimuli caused by photons that have interacted with the object. Again, I think that this shows faults in the claim of the non-reality of quasi-particles.

  2. Thanks for mentioning my paper, Ross, and thanks for alerting me to your blog post. Actually, I don't feel all that strongly about whether one should refer to quasi-particles as 'real' or not -- in some sense, they are quite obviously 'real enough' (that is, 'manipulable'). The main thrust of my paper is to show that Hacking's criterion of manipulative success alone cannot provide an answer. Recall that Hacking is willing to bite the bullet on gravitational lenses, which he regards as 'unreal' (even though they can be observed, if not with the naked eye then with fairly straightforward detection devices such as telescopes) and which fail to be manipulable for the purely contingent reason that they are huge objects far away from Earth. Surely, excluding gravitational lenses ('clusters of big heavy objects') but including quasi-particles ('collective effects of many-body systems') shows that the criterion of manipulative success falls foul of double standards. However, there is a deeper point, too. As Quine famously said, 'no entity without identity'. This seems to guide the entity realist's position. But quasi-particles are parasitic upon an indeterminate (possibly varying) number of particles, and in this sense lack determinate identity conditions. So, in this sense, it is importantly *not* just about observability or detectability, but really about which entities are natural kinds and which are derivative. Of course I am not disputing that quasi-particles are real /insofar as/ they are collective effects of a many-body system -- but that's all they are. (By the same token I find that the share value example actually works in my favour. If someone were to suggest that there is something metaphysically 'deeper' to the existence of share values, over and above the fact that market participants are, on average, willing to trade certain stocks at a certain price, than I think he/she would be committed to a rather odd world view.) I think what we should say instead is that questions of scientific realism cannot be answered entirely without reliance on theoretical background knowledge.

  3. How have you come to the conclusion that electrons and protons are "real".
    What's your definition of real?
    What definitions of electrons and protons are you using?
    What definition of quasi-particles are you using?

    naturally by changing the used definitions of the components you can arrive at pretty much any conclusion and indeed there are in a less strict mathematical sense an infinity of usable definitions for each.

  4. Interesting stuff.

    I've been reading X.-G.Wen on QFT and he likes the idea of a string-net condensation's collective excitation as the source of electrons and photons (the photons are vibration modes of the strings and the electrons are the ends?) I also recently read that the forces weren't distinct until a few instants after the big bang. It makes me wonder how I can claim that anything is more real than anything else.

    Not to get too far outside my field of expertise, but happiness is real despite being a collective chemical excitation (is that splitting the infinitive?)

    And phonons are real because they have such a tidy propagator.

    All very different uses of the word real though. Perhaps it is just semantics as I think Vincent implied.

    Is a stock price real if nobody pays it? Just kidding...