The concept of operational parameters is discussed. Operational parameters are compared with extrinsic or hidden variables. It is shown, that by applying Einstein's synchronization conventions, the transformation properties of space and time parameters depend on the dispersion (energymomentum) relation of the system. In particular, with nonlinear forces acting in a medium, transformation laws need not necessarily be of the Lorentz form.
Among other aspects, there are two principal ways of describing a system: the operational [or intrinsic] parameter description, and the extrinsic parameter description. Heuristically speaking, the operational parameter description may include every parameter which can actually be measured from within the system in the sense that devices can be constructed and processes can be imagined that would lead to a measurement of such a system parameter. This would eliminate idealistic concepts or mere hopes that it would eventually be possible to find the ``true and real'' value of a system parameter with an associated uncertainty zero. Such illusionary ways of description are referred to as extrinsic, since it would require a point outside the world, from which it is possible to look at the world, without disturbing it at all. Its parameters may be hidden parameters from within the system. Since early Greek philosophers such as Plato and Archimedes^{1} recorded such considerations, it is quite obvious that natural sciences have to rely exclusively on operational parameter descriptions. More recently, the use of operational parameters has been stressed for statistical physics^{27}. However, to my knowledge, no comprehensive study on these notions has been undertaken yet. The operational parameter description is the only physically meaningful representation of a system, and it has many advantages. One of these advantages is its pragmatic approach: in particular, the observer [the measurement apparatus] need not and cannot be excluded from the observed system. The operational parameter description needs no knowledge about the closedness of a system to its environment, since it does not attempt a comprehensive, inclusive description of that particular system. This ignorance is indeed inherent for any parameter description, since one never knows for sure the forces acting in a system, nor will it ever be possible to project future developments of knowledge, which may eventually yield corrections within the range of present uncertainty. There seems to be an unavoidable [psychological] consequence: the acceptance of the preliminarity of all scientific findings drives investigations to consider all possible operational descriptions and their underlying structure. There cannot be anything like a preferred parameter description for other reasons than subjective ones.
In section II a very brief introduction to the operational parameter description is given. It is shown how a sort of quasiextrinsic parameter description can be developed. In section III these concepts are applied to spacetime parameters in the field theoretic description of quantized media. It is shown that their transformation properties depend on synchronization conventions and on the dispersion [energymomentum] relation in a particular medium. In the last section IV speculations about a Lorentzinvariant vacuum medium and experiments for its detection are introduced.
For clarity and consistency, the operational parameter desciption is introduced. Assume a system S. Then an operational parameter description P(S) = {S;p_{1},...,p_{n},...} with parameters p_{i} [it is assumed that the set P(S) is countable, but this is not necessary] contains parameters which could at least in principle be measured by devices and processes available in that system. Following Einstein^{9} it can be said that each element of P(S) corresponds to an element of reality .
Such a parameter description may be consistent and unambiguous [but it need not be] if both the system and the devices and processes are consistent and unambiguous. Since this is not a necessary condition, operational descriptions are quite arbitrary and have to be selected by economic, symmetric and progressive reasons [this introduces a subjective element to a selection procedure]. Here progressive means that the description is able to foster future developments in that area, as has been intensively discussed by Lakatos^{8}. It is only a belief that the most economic and symmetric parameter description is the most progressive one. As historic examples show [for instance the Ptolemean versus the Kopernican system], this need not always be the case.
Next I shall try to construct examples of operational parameter descriptions and introduce the concept of quasiextrinsic [almost extrinsic] descriptions. Let us consider two systems S_{1} and S_{2}, an interaction I acting in both of them, and two associated operational parameter descriptions P(S_{1},I) and P(S_{2},I). We shall define the system S_{2} approximately closed with respect to S_{1} and to the interaction I, if S_{2} responds only ``slightly'' to changes in S_{1}. Formally, this situation can be characterized by:

A parameter description is called independent , when all parameters commute with each other with respect to the Poisson bracket:

A parameter description is called complete , if no parameters p exist such that p is independent with respect to P(S,I), but does not belong to it:

Generally, if the parameter descriptions P(S) and P(S,I) are complete and independent,

Assume two systems S_{1} and S_{2}, and two interactions I_{1} and I_{2}. Assume further that S_{1} and S_{2} are approximately closed with respect to one interaction, say I_{1}:


In this context, the extrinsic parameter description might be defined via a limit: a parameter description P(S_{1},I_{2}) from S_{2} is called extrinsic, if S_{1} and S_{2} are totally closed with respect to both interactions I_{1} and I_{2}. Clearly, this is impossible to realize, since there cannot be any exchange between systems without altering the states of both.
In what follows I shall give an example of such a configuration: assume a pool filled with water, which will serve as system S_{1}. Let us assume further an optical instrument recording electromagnetic radiation as part of our system S_{2}, and the interactions I_{1} and I_{2}, being identified with water wave interactions and electromagnetic interaction respectively. Since light does not affect water wave dynamics appreciably, but changes the state of the optical instrument, a realization of the described cybernetic model is obtained, with the optical instrument yielding a quasiextrinsic view of the pool.
Now suppose there is a waterflea [Cladocera] living on the surface of the water. Suppose further that this creature is blind, i.e. it is not able to employ electromagnetic radiation. Then its operational parameter description will almost certainly purely have to rely on water wave dynamics as operational device. On the other hand, another creature [such as a bird], seeing the surface of the water by light, will be able to choose between the water wave dynamical description by electromagnetic scanning [i.e. photography or that like] and wave scattering; both are operational for this creature. Whereas the waterflea, if it is not imaginative enough, will always wonder about the form of its world and its embedding and limitations, the air creature will immediately overlook the situation, thereby rather using its eyes than produce wave scattering. That of course does not mean that the waterflea in principle can never detect light; its just that electromagnetic effects yield such small contributions to phenomenology that they are difficult to detect.
For example, the most obvious and evident operational parameter description with water waves is by a sort of quantisation with the help of Poisson's brackets^{11}

The question arises, what is the use of such a water wave description ? One answer is certainly pedagogical; we know it better: from our quasiextrinsic point of view, water is just made out of molecules forming waves. We stick to what appears most evident to us  but not so for the water flea! So, probably we can learn from this example that we have to be very careful by using parameter descriptions and relying completely on their conclusiveness and uniqueness.
As an example I shall try to answer the question, which extrinsic parameter description might be used for space, time and energy in a quantized medium.
So far nothing has been said about the formal structure of theories corresponding to parameter descriptions P. As a postulate it is assumed that any theoretical structure can be characterized^{13} by an algorithm A, encoded by a string formed by some alphabeth a. It could be expected that there is a onetoone [bijective] correspondence P º A. In this case we may say that A is inclusive corresponding to P. However, this saying is probably of not much use, since one never knows if P is an inclusive representation of the empirical data or reality [or if one does not accept realism in this respect: one never knows if to a later time there will not be a more inclusive P¢ of the same system, such that P(S) Ì P¢(S)]. The only qualitative correspondence between P and A seems to be a historic one: to every historic moment, a temporal correspondence of some representation of the world P with some explanation of the world A is attempted.
Whether algorithms can be accepted which do not contain all empirical content, such that P É A, or whether they contain hidden variables such that P Ì A or a mixture of both cases for different elements of reality in the sense of Einstein^{9} is quite arbitrary. If these algorithms are reducible [such that a shortcut is possible to obtain the parameter description], or NPcomplete [such that they can be solved with an arbitrary number of parallel computational paths and in polynomial time but are computational irreducible], or PSPACcomplete [such that they can be solved with polynomial storage capacity, but require exponential time], are still open questions which have been edged on only recently^{14}. I shall not pursue them here.
As an example I shall discuss now a theory of a quantum medium which may yield the wellknown covariant transformation laws of space, time and energy in some operational parameter description, but is certainly not manifestly covariant from an extrinsic or quasiextrinsic point of view. It can be seen that certain symmetries such as general covariance emerging in specific parameter descriptions, may not be evident for others.
For further considerations it is assumed that it is at least in principle possible to construct measurement apparata for certain observables and parameters such as space, time or internal quantum numbers out of elementary processes in a system. In that way, frames of reference can be defined in an operational way.
In what follows a time scale will be defined operationally. A model atomic clock is constructed which is at rest with respect to the rest of the system. A time t_{A} and t_{B} is assigned to its initial and final states \mid yt_{A}ñ and \mid yt_{B} ñ . Both t_{A} and t_{B} may be arbitrary real numbers, with t_{A} < t_{B}. It is assumed that the state of the clock can be expanded into a sum of orthonormal state vectors \mid Iñ and \mid IIñ:




Having defined intrinsic scales for clocks at rest with respect to a given system or frame of reference, the attention is drawn to the relation of scales in different frames of reference. Again, an operational approach is chosen by considering two clocks, one at rest with respect to the frame S_{1} and the other with respect to a second frame S_{2}. In order to compare scales, a transition of the state of the clocks that can be measured from both systems has to be considered: assume a single clock, say the clock at rest in S_{2}, described as above by the amplitudes áy_{1}t_{1B}\midy_{1}t_{1A}ñ in S_{1} and áy_{2}t_{2B}\midy_{2}t_{2B}ñ in S_{2}. As the square of these amplitudes has a probabilistic interpretation and by identifying \mid I_{2}ñ = \mid I_{1}ñ = \mid Iñ and \mid II_{2}ñ = \mid II_{1}ñ = \mid IIñ, the following identities are obtained:





The clock under consideration, that is the one at rest relative to the second frame S_{2}, moves with a velocity v_{1} (measured in S_{1}) relative to S_{1}. The frame S_{1} itself is at rest relative to the medium. In this configuration, w_{1} can be identified with the dispersion relation


However, if it is not easy or impossible to distinguish between frames of reference by measuring their motion relative to the medium, the value w(p = 0) for zero velocity may be assigned to w_{2}. More precisely, this can be formulated in the following way: if an object at rest in the frame of reference S_{1} with energy e_{1} is transferred to another frame of reference S_{2} so that it is at rest there, an energy


Generally, dispersion relations may depend not only on momentum, but for inhomogenuous media also on space coordinates of arbitrary dimensions and for nonlinear forces on the square of the amplitudes as well as on other parameters such as a fundamental length. In these cases considerations are quite subtle, since the sound velocity may depend on the frequency, and hence, scales would depend on the frequency chosen for synchronization.
One of the more interesting systems and one which has been extensively discussed in the literature^{17} is a threedimensional continuous,homogenuous and isotropic field with its points or constituents coupled linearly to the equilibrium position and neighbors. It will serve as a perfect medium model, reproducing Lorentztype transformation laws. There is a wide class of physical problems, where the medium can be linearized in that way: for example, acoustic phonons and other quasiparticles in solid state and manyparticle physics. The dispersion relation is given by w(p) = [(cp/k)^{2}+w_{0}^{2}]^{1/2}, where c and w_{0} are two constants proportional to spring strengths and k is the minimal action of (2.7). Again a configuration is considered in which a clock travels with velocity v_{1} = dw_{1}(p)/dp, with respect to the medium. In this case it is possible to identify w_{1}(p) with w(p), and the following time dilatation is obtained:


If the vacuum is thought of in more realistic terms as a ground state of a quantum medium, then the immediate question is: what can be the use of such a notion ? Despite other more subjective reasons the ultimate goal should be the experiental detection of this medium, which might subsequently yield possibilities of transforming the vacuum into new phases with very different characteristics [maximum propagation speed higher than the speed of light, higher action resolution and that like]. Although these are mere speculations at present, they represent a progressive scientific program, which has to be tested experimentally^{8}.
This rather utopian goal has to be initiated by a careful study of possible indicators for such a medium. In what follows I shall give a brief list of operational parameters which have the potentiality to indicate a medium:
(i) granular structure of spacetime : its detection will very much depend on the type of granularity involved. The simplest case of an energymomentum cutoff corresponding to a fundamental length has recently been discussed by Shupe^{18}. But much more complex and subtle irregularities may occur^{19,20} yielding a fractal spacetime with associated Hausdorff dimension not equal four^{21}. Since this has an impact on all quantum mechanical matrix elements, in particular sensitive radiative corrections such as the anomalous magnetic moment, this noninteger Hausdorff dimension would operationally be attainable: with present accuracy, it is 4(5,3±2,5)×10^{7};
(ii) the cosmic background radiation represents the possibility of the existence of an absolute frame of reference, but it is not a forcing reason to accept a quantum medium, since this only concerns the initial condition of the world rather than the structure of the theory and the associated equations of motion. However, if a quantum medium should eventually be detected, it will be presented as the first decisive indication in future physics textbooks;
(iii)deviations from the theory of special relativity: since its introduction the theory of special relativity has developed into a reference theory, almost sacrosanct to objections. Despite prevailing discussions about the detectability of the oneway velocity of light, there is no evidence of failure. However, some deviations from the ideal dispersion relation E(p) = [(cp)^{2}+(m_{0}c^{2})^{2}]^{1/2} due to nonlinearities of the quantum medium may eventually occur at very high energies, but there is no scale to tell what high in this respect means. For a detailed analysis of the present experimental tests of special relativity and trends, the reader is referred to Refs. 2224.
(iv) there may exist a large number of further effects that have not been proposed yet, but will shed new insight into the theory of the vacuum structure. Presently it seems unlikely that this will happen soon, but one never knows. In this context experiments to test fundamental properties of spacetime and quantum structure in extreme limits are necessary.
In conclusion it can be said that since we are not able to look at our world from above and utilize an extrinsic parameter description, we are forced to use operational parameter descriptions [presently, there is not even a quasiextrinsic parameter description conceivable]. It has been shown that from a phenomenological point of view [with present accuracy], a medium with a particular type of dispersion relation yields an operational parameter description for spacetime and energy identical to parameter descriptions obtained from the abstract assumption of Lorentzcovariance. This fact alone would not justify attention. However, the Lorentzinvariant quantum medium yields a phenomenology not predictable by the special theory of relativity. \bigcirc
I am indebted A.Zeilinger for many helpful discussions, as well as G.Grössing for pointing out the importance of the cybernetic approach. This work was supported by the Austrian Ministry for Science and Research, project number 19.153/326/85.
B.Russel,A History of Western Philosophy ,Allen and Unwin, London 1964 B.Misra, Proc.Nat.Acad.Sc.,U.S.A.,75(1978) B.Misra,I. Prigogine and M.Courbage, Physica 98A,1(1979) I.Prigogine, Science 210,777(1984) C.M.Lockhart,Ph.D. thesis, University of Texas at Austin 1981 C.Lockhart,B.Misra and I.Prigogine, Phys.Rev. D25,921(1982) H.Narnhofer,M.Requart and W.Thirring, Commun.Math.Phys. 92,247(1983) I.Lakatos, The methology of scientific research programs, Philadelphia Papers, Vol. 1, J.Worrall and G. Currie (Eds.), Cambridge University Press, Cambridge 1978 A.Einstein, B.Podolsky and N. Rosen, Phys.Rev. 47,777(1953) F.Varela, The Principles of Biological Autonomy,North Holland, New York 1980 H. Goldstein, Classical Mechanics, AddisonWesley, Reading,Mass. 1959 F.S. Crawford, Berkeley Physics Course, Volume 3 (Waves) ,McGrawHill, New York 1965, sects. 6.26.3 J.E.Hopcroft and J.D.Ullman,Introduction to Automata Theory, Languages, and Computation,AddisonWesley,Reading,Mass. 1979 Wolfram, Phys.Rev.Lett.54,735(1985) A.Einstein, Ann.Phys. 17,891(1905) R.Mansouri and R.U.Sexl, Ge.Relativ.Gravit. 8,497(1977) E.M.Henley and W.Thirring, Elementary Quantum Field Theory,McGrawHill,New York 1962 M.A. Shupe, Am.J.Phys. 53,122(1985) P.C.W. Davies, in Quantum Gravity 2, ed.by C.J. Isham et al, Cladendon Press,Oxford 1981, p.207 C.J. Isham, in Quantum Theory of Gravity, ed.by St.M. Christensen, Adam Hilger Ltd.,Bristol A.Zeilinger and K.Svozil, Phys. Rev. Lett. 54, 2553 (1985) A.K.A.Maciel and J. Tiomno, Phys. Rev. Lett. 55, 143 (1985) E.Fischbach et al., Phys. Rev. D32, 154 (1985) D.W.MacArthur, Phys. Rev. A33, 1 (1986).