Essential Structure in Physical Observation.

Essential Structure in Physical Observation
Working paper

A.G.Booth	Original 3 March 2002	Copyright © A.G.Booth 2002-2004 All rights reserved
Document ident:	Last updated 22 Dec 2006	Essential Structure in Physical Observation (work). A.G.Booth
Keys:	cybernetic physical observation observer emergence engineering quantize quantization quantal quantized entangled quantum entanglement attractor

Introduction
Semantics
"Probe" and "Respondent"
Frames of Reference and the Probe Part
Amplification and the Respondent Part
Thermodynamics
Fluids
Resolution into Discrete Ensembles
The Question of Unilateral Amplification
Kinds of Amplifiers
Some Definitions and Lemmas
Coupled Observers
Cybernetics Rationale for Quantisation
Notes
References and Background Reading

This symbol denotes linkage to an independent web site:

Introduction

At the heart of any process of observation there are certain basic things which must happen. These might be analysed in abstract terms, or then again in terms of the quantum physical irreducibles. Here I attempt to link between these concerns with the aim of clarifying a path to the quantum mechanical form of model, and at the same time to provide the basis of a macroscopic model.

A discussion of the motivation for this paper can be found at "Reconciling Physics with Cybernetics." ref [ABo00] and "Why Seek Frequency Domain Continuum Models?" ref [ABo01].

A link to relevant cybernetics concepts (Ernst von Glasersfeld and Humberto Maturana) can be found at "Distinguishing the Observer: An Attempt at Interpreting Maturana." ref [EvG97].

Parts of the paper contain items which are conjectural or untested. Because of this they are liable to editing changes from time to time in the web published copy. The "Document ident" shown just below the header of the paper and showing date of last update should be used to check for the specific form of the paper, in cases where that might matter. Please communicate comments to Tony Booth.

Semantics

There are some substantially separable aspects to the meaning of the word "observer". One of these is concerned with the point of view or frame of reference which is involved in the observation and another is connected with the information or data which is gathered by and subsequently made accessible by this sub-system referred to as an observer. There is also another aspect involved which might be regarded as forming a "point of view", and that is the way that the data is interpreted before it is finally expressed, but that issue will not be discussed explicitly here.

There is an implication within the meaning of the word "observer" that information may be gathered which is absolute in the sense of being objective. There is thus a curious paradox in the usage of the word because the sense of point of view of observer is strong in its meaning and contradicts the idea that the result of the data gathering operation is in some sense absolute or objective. Splitting apart the functions within observation as is done here offers clarification and a path round this paradox.

"Probe" and "Respondent"

As an alternative we might consider a word something like "respondent" for use where the ability to enter into communication or dialogue following observation is of dominant importance. This would then leave the words "observation" and "observer" to be more concerned with the frame of reference aspects than with the informational content aspects. The word "probe" is also sometimes used (e.g. see [BK92]) to refer to the part of a system of observation which is in direct contact with the observed sub-system. Thus a complete system of observation might be divided between a probe part and a respondent part. It is interesting to note that the word "respondent" links usefully with the word "correspondent" as the concern over communication changes from initial results of observation to dialogue about earlier obtained data.

The Basic Structure of an Observer

Frames of Reference and the Probe Part

The frame of reference or point of view of an observation is determined by the placement of the probe part of the observer structure as it contacts the observed system. The issue of frame of reference applies in widely different sets of circumstances for instance from physical systems where metrical dimensional frames of reference might be the main concerns through to social and political systems where qualitative aspects are the usual concerns. An interesting simple example of the latter is the one dimensional axis between liberal and conservative thought. On this axis a position is often considered or discussed and relative statements and comparisons are also frequently made. By contrast, if a statement is made in a multi-dimensional political framework, such as that obtaining in a plural interacting group situation, then that statement is likely to be nearly impossible to interpret within the one dimensional framework of the liberal/conservative scale.

For further elucidation we may study as example a physical process in which our human perception of space is more elaborate than that in which the example system is considered to perform its development, observation and dialogues. A simple case of such a system is that of the mobility of certain bacteria in the presence of a gradient of beneficial potential in their chemical world (see "The Cell" pp720-721 [BA89]). These organisms, using only a single dimensional time related response (apparently having zero spatial dimensions in their model) and working with only a single actuation process through the reversal timings of a helical propulsion flagellum, are capable of controlling their position to gain advantage in a gradient of beneficial potential in a fluid. We may use this example to note how a system with no explicit coding for a three dimensional navigational problem can usefully perform such navigation.

It behoves us to conceive that we may be in general in a like position with regard to higher dimensional problems. Indeed we are practically bound to be in such a situation regarding our manoeuvres in the many conceivable sorts of spaces in which we see ourselves as existing, these spaces being either recognised by us or not.

Amplification and the Respondent Part

Without concern for the difficulties with frames of reference we may set about describing the nature and structure of a sub-system which behaves as a "respondent". This is concerned with that part of the overall operation of an observer which makes it possible for communication of gathered information to occur as the outcome. An important feature of this lies in the necessity for certain energy conditions and circumstances to exist for communication following observation to be possible at all. In particular there is a need for the process to contain some sort of amplification of the signal.

There are two senses involved in the use of the word "response". The first of these is that a signal might be stored for substantial time following observation (this we can call "recording") and the second is that a signal might be produced following enquiry (call this "expression"). In the end both might be of interest, but generally these two senses of the word "response" can be kept separate. In any case there is a universal requirement for amplification to be present either prior to or within each of these two aspects of response.

Further, regarding this amplification, it is necessary that it shall be to some degree unilateral. This means that the attenuation of reverse flowing signals shall be greater than the amplification of forward flowing ones. There exist certain simple sorts of amplifier structures all the way down to the quantal levels of mechanism which cannot satisfy this requirement. It is necessary for there to be a mechanism with non-reciprocal nature in order to achieve this effect of unilateral amplification. Note that any system defined by a constant unitary state evolution operator (i.e. all the conventionally considered quantum mechanical evolution operator models) cannot develop this essential non-reciprocity. It is only by virtue of non-linear operation in causal loops (through reflexive action of the state to modify the state evolution operator) that it becomes possible.

If such unilateral amplification is not present the processes of replication and propagation of the amplified signal will reflect disturbances into the original observed state at least in proportion to the amount of output power delivered. In particular, any echo of signal arriving at the amplified output point of the system will return to disturb the original observed state by more than the inherent dissipation involved in the essential action of sampling. Macroscopic interactions at the output point would also induce similar significant disturbance at the original state variable.

Thermodynamics

We cannot avoid accepting that the sort of system in which we are interested is in a fundamental sense thermodynamic. This will include living systems. A system containing an observer is thermodynamic because it is essential that its information shall be reproducible. Because of this it must be possible within what is defined as the system for a small and local condition to influence large and relatively remote states of things in (at least approximately) a consistent and predictable way. Determinate schemes of description cannot allow this feature of a system because the overall determinacy is the only influence available. Thus if we choose to use a determinate model of local underlying process it is inherently necessary to deal with the distinction between this process and that of any other reality, namely, our "thermodynamic system".

In so far as such a thermodynamic system involves quantum mechanical effects, consider for instance the storage of information in molecular structures in DNA, it is still essential that it shall everywhere maintain sufficient immunity to corruption by thermal or other unpredictable interferences. This protection of information from corruption consists in many different guises of a similar mechanism, namely keeping the likelihood of significant interference by thermal or other means sufficiently small to maintain adequate probability of survival of the overall system. Herein lies the necessity for what we can call "state attractors". If a disturbance occurs the state must tend to settle back to the nearby preferred value. In order to be stable in this way these preferred attractor values must be pre-ordained, so it is inherently necessary that the states in which the system can be found must display a quantal or discrete nature.

In quantum mechanical terms the requirement is to keep a segregated ensemble of fermionic entities (say a group of electronic wave modes) as distinguishable regarding its state vis-à-vis its alternative states. This must be done both where the problem is one of storage of information and of its transformation. In the case of transmission it can also apply where a fermionic carrier (say a free electron or ion) is itself transmitted, but has a somewhat different form when electromagnetic field is the means of transmission.

It is currently commonplace in the physics literature to maintain the quantal model throughout the electromagnetic parts of the process too. Unfortunately that has an effect of making the entire system difficult to handle at the interface between the quantal and the thermal -- it all appears to be quantal! It is a reductionist determinist model, and as such does not have any terms in which it can posit the boundary of the living entity, or the interface between observer and observed sub-systems. It presents difficulties of comprehension of the conceptual distinction between the thermodynamic and the underlying process. It is up to the model builder to define notionally where is the boundary between observer and observed as two parts of a physical system, and the resulting alternative models must be consistent in the way they handle this distinction.

The situation is further complicated by the possibility that we may choose to allow that the underlying process is itself subject to modelling by probabilities amongst intractably large numbers of entities. It will itself then be thermodynamic in nature too.

Fluids

Connected with any amplification or information reproduction process is an interaction of a fluid with a structure. The fluid is capable of taking different states, either by its quality or by its quantity within a given segregated space. The structure must mediate the motion of the fluid. The aggregates of the fluid must be capable of modifying the structure.

For this to be possible there must be other structure which mediates the interaction of the fluid aggregates with the first structure.

The concept of fluid starts with a sense of conservation and a concomitant sense of continuity (departure from one space necessarily implies arrival at some other). However a fluid can also carry other attributes; perhaps it is even possible to base the notion of fluid upon these non-conserved attributes without any conservative aspect. An example of this would be a smell pervading a building space as a fluid … there is no particular sense of conservation. By comparison the sense in which air pervades such a space does have conservation aspects … such as mass, total heat, and perhaps others. There are other difficulties, for example the total heat of a given sample of gas might be changed without its constitutive aspects (e.g. mass and identity) changing, in which case the heat is itself really the fluid, not the gas which is carrying it.

Given these difficulties with the concept of “fluid” it might be better to avoid it in the definition of amplification and instead use a more abstract conserved “stuff”. Energy would be one of these (a pity it fails to be relativistically invariant!). Entropy can only remain conserved if the framework of its definition remains constant in the observer of it … it has a relative nature.

Resolution into Discrete Ensembles

Segregated ensembles or packets of carriers can be used in many macroscopic ways as means for storage of information, but at the ultimately small level of physical process the carriers can be thought of as atomic structures and their discrete states of ionic coupling. Optical quanta can be considered as carriers, but when this is done the problem remains one of calculating with probabilities somewhere else. We shall here continue to treat the electromagnetic field as the non-quantised wave function which mediates segregation or qualification of states of quantally ionisable entities. In particular the interactions between electromagnetic field and electron wave field can in many cases be viewed as central in bringing about changes of structure or ionisation.

The Question of Unilateral Amplification

When we discuss amplification here as being "unilateral", this does not mean that there is necessarily no reverse directed causal influence at all, but only that there is a coupling which has forward power gain and for which the product of the forward and reverse causal power gains is less than unity. Such amplification cannot be realised by a system with constant unitary state evolution operator. It therefore follows that non-linearity in the form of state feedback modulating the structure of the state operator has to be present.

One (and definitely not the only conceivable) possible manifestation of the effect of this sort of non-linear feedback is what we observe in nature and refer to as "charge quantisation", and the element quantum of charge is observed to be a constant. Since at the ultimate microscopic level this appears to be uniquely the form of the nonlinear manifestations in our universal environment, we can proclaim that these phenomena must be involved in every real unilateral amplification process.

Because of the fundamental importance of these effects in observation it is therefore interesting to seek to comprehend ways in which the primary models of physical processes, such as are based on Klein-Gordon, Schrödinger or Dirac equations, can through their non-linearities give rise to charge quantisation. Maxwell's equations can also be involved, but they do not themselves introduce any non-linearity.

Mechanisms which provide non-reciprocal coupling regardless of or without the involvement of amplification are conceptually of interest here. The most famous form of such a mechanism is the Faraday rotation effect. These processes arise from interactions between an electromagnetic wave and a mass of ferromagnetic matter with its magnetisation aligned with the direction of propagation of the electromagnetic wave. It causes the transverse axis of polarisation of the electromagnetic wave to rotate as it progresses along the direction of wave propagation. By combining a process of this type with an amplification process which is not unilateral we can produce overall a unilateral amplification process. Perhaps this is not such a surprise when it is recognised that magnetic field does itself display quantal phenomena. Our need is probably just to comprehend in suitable terms the nonlinear/attractor mechanism underlying this form of quantisation too.

* Mixer Regime.

A detection structure can be put together by aggregating the output of a "mixer" downconvertor with energy collection in an atomic medium excitable at the reduced output frequency. No particulate phenomena appear explicitly with this process. There is only a noise which can be associated with the inherent phase uncertainty of the mixing medium state recharging process which is not necessarily seen as quantised. In processes of this type all summation of effects used to achieve macroscopic output state is incoherent.

Reciprocity extends into the frequency shift domain. The problem of achieving unilateral amplification arises here too.

* Avalanche Regime.

In avalanche amplifiers the mechanism of energy growth in amplification is conceived as separate from the probe coupling of original detection (as it might have been without the amplifier structure present). However when coupled together these two processes cannot operate separately. That is an idea with no real meaning. It is the amplifier dynamics with its inherent fluctuating incipient instability which determines the eigenvalues of the dominant activity when avalanche occurs. These perturbational effects with their associated transient modes and eigenvalues operate coherently throughout both the amplification cascade and the probed region under observation.

Any possible structure of detection which can produce time related information of the "particle detection" sort uses avalanche amplification. This avalanche can have divergence time (much!) shorter than the reciprocal bandwidth of frequency domain diffusion of the (hypothetically unobserved) process under observation. This leads to a spurious sense of short term correlation through its apparent involvement of fast processes. These short term effects may be validly interpreted as artefacts of the coherent amplification process rather than any sign of particles being present.

The only possible symptom of particulate process being present lies in anomalous departure of the approximately Gaussian distribution of pulse heights away from the form expected due to the upward thermally fluctuating mode energies making excursions into the conditions supporting the avalanche process. If the width of this distribution is less than expected from the dynamics of the underlying diffusion process believed to be present then this might be usefully interpreted as particulate activity. However, the presence of nonlinearity of the attractor type in the state mechanism of the probe region of the process can also produce this effect, so the argument for particles being necessarily present would then still not be complete!

When the coherently conjoined processes include two detectors with independent amplification cascades then the artificially short times of correlated strong activity throughout the system (including both detectors) produce an effect of simultaneous action at a distance and this can appear to involve non-local action. Cases of this effect include:

"Photon anti bunching" -- fast non-local negative correlation as observed when the coupling modes are filtered through a single trapped atom, see H.Schadwinkel et al [HSc00] and also V.Gomer & B.Ueberholz [VGo]
"EPR paradox" -- fast non-local positive correlation as observed when a coherently coupled pair of Rabi processes link orthogonally to otherwise independent detectors, see Michel Janssen [MJ00]

These effects are subject to the prevailing parameters of the observation structure (consider polarisation parameters in the Freedman-Clauser experiment) and may be averaged throughout the period over which the detection effect is sought or over the diffusion decorrelation time of the underlying probe process, which ever is the shorter. In those interesting cases which take on the appearance of non-local effects these averaging periods would need to be, and indeed appear to be, substantially longer than the avalanche divergence period.

Kinds of Amplifiers

* Thermal continuous.

This includes such as transistors, thermionic vacuum devices, reverse biased photo diodes. Excitation is applied to the post-quantal (resolution) species ... The dynode (electron multiplier) is in this class, though when cascaded to make an electron multiplier the alternating stages of probe and resolution species produce a quantal effect in an early stage.

* Thermal quantal.

This includes avalanche photo diodes, ionisation track chambers, Geiger-Mueller tubes. Excitation is applied to the pre-quantal (probe) species ...

The dynode (electron multiplier) would usually be assigned to this class of amplifier, and in a cascade that appears to be valid. However see the note above under thermal continuous amplifiers.

* Stimulated emission.

This includes such as lasing amplifiers. Excited species contribute coherently in a process which is entirely pre-quantal. Noise source is spontaneous emission in the thermal equilibrating process of the pumped species. Through the nature of atomic resonances (over which this equilibrium is being maintained as a Fermi-Dirac distribution of excitation levels) this leads to an equivalent lower bound to noise as for quantal processes, but it is usually best expressed over a space-frequency continuum. Without additional arrangements (e.g. Faraday isolator) these amplifiers are not unilateral.

* Negative resistance.

A notable form of negative resistance device is the tunnel diode electronic circuit element in which the current will decrease as the potential difference across the diode is increased over a certain part of its range. With it an amplifier can be built, but alone it cannot be unilateral.

* Parametric amplifiers.

A remarkable process in its ability to amplify from a noise level substantially below that equivalent to the temperature of the electrical support matrix. This needs to be compared with the corresponding capability of some sorts of thermal continuous amplifiers. As for the case of stimulated emission amplifiers this process is not, of itself, unilateral.

* Rotating field.

The basic structure needs to be developed here.

Some Definitions and Lemmas

Lemma 1: There is no (useful) observation without unilateral amplification.

Definition of "(useful) observation".
Observation (of the useful variety) is the process whereby a sub-system called an "observer" changes state in such a way that repetition of the process independently but under similar conditions either by another such observer and/or at another time can (be claimed to) show a significant correlation between the resulting changes of state.

Definition of "amplification".
Amplification is the absorption (by dissipative observation of it) of a signal represented in some given set of elementary physical (i.e. unitarily couplable) variables and delivering a new representation in another set of such variables with reduced dissipative sensitivity to its further observation. In other words, the process of amplification creates a calculable amount of assessable degradation of some given signal and yields another signal upon which further observation to the same degree of closeness produces less assessed degradation.

The second law of thermodynamics tells us that:

Any process for performing amplification produces, however small, some increase in signal entropy.
Any process of performing amplification consumes fuel, i.e. reduces the rarity of state in some species supplied in a rare state compared to the bulk of that species in a common state.

For the definition here, even the mere purpose of establishing a theory is sufficient to be regarded as "useful" in the above sense of observation. Otherwise the theory could not be verified.

Lemma 2: There is no unilateral amplification without associated enduring changes of ionic states.

Note: There is a point of contention regarding this lemma. There exist processes of amplification in which the unilateral feature is realised by a field interaction between the signal as an electromagnetic wave and spin/magnetic processes. The most famous manifestation is the Faraday effect. I have not yet checked sufficiently the validity of the lemma for these sorts of amplifiers as would suit the context of this discussion. The conventional quantum mechanical analysis is not necessarily satisfactory, though the magnetic processes are subject to quantisation phenomena on their own account. -- Work ongoing.

Definition of "(enduring) change of ionic state".
Ionic state (of the essentially enduring variety) must display state attractor behaviour. This means that under disturbances less than some magnitude they must tend to recover the prior state. They must not accrue marginal changes as a result of such weak disturbances.

There are no structures simpler than nucleons (including baryons and the more complex general case of hadrons) which can have enduring ionic state associated with them. Isolated electrons and photons, though they can have vectorial energy, display no attractor behaviour and so cannot carry enduring ionic state except insofar as they terminate by absorption into an atomic structure (i.e. involving hadrons). There also appear observable entities with short lives which can exist in isolation and that during their life can display ionisable properties (the heavy leptons and pions), so there may be some special cases of short but significant duration ionic effects related to those.

If the changes of ionic states are confined to local molecular structures then the process is one of chemical recording; film photography works like that. If the media of the ionic state changes are abstracted away in a systematic manner then active systems with fixed structure become possible. Electronics does this by carrying the results of ionic state changes away in the form of electric current, and biological hormonal systems do it by using alternative states and concentrations of mobile molecules as means of carrying information.

Lemma 3: There are no observable changes of ionic states without quantal phenomena.
This is because states correspond to the selection of attractor points near to which a system operates, and that changes of such selection can only occur in discrete steps because the attractor points are discrete.

Although such step changes of attractor selection are discrete they are not instantaneous (though particle theories tempt many people to think of them as so). Depending upon the form of the disturbances and system structure there can be many sorts of transient states during changes of attractor selection.

The above lemmas are valid in a pure wave model as well as in a wave mediated particle model.

So long as these quantal phenomena derived from a wave model can account for the results of observation there is no absolute necessity to presume the existence of particles. Indeed in such a case there is a substantial risk that their introduction as a concept will over specify the model, and thereby create expectations of behaviour which are beyond and unrelated to the range of observable phenomena. If not needed they may complexify the model rather than simplify it.

Because of the above, the justification for using particles as part of an otherwise undeniably wave mediated model must then be based upon evidence that the model structure is made simpler by the introduction of those particles.

So long as the domain of concern is closer to the results of observation then the particles do indeed simplify things because they correspond more or less to the discrete momentum conserving entities apparent in the observed results (hence the suitability of Lagrangian dynamics at this point). However when the domain of concern is followed to the minimisation of the underlying model or to the nature of the underlying processes including observation itself then these particles can become a burden of over specification. To deal with this in the first instance (the current convention) they can be treated as differential elements and integrated out of the problem over a (so called "probability") space of simultaneous existence (e.g. integration based upon structures depicted in Feynman diagrams). However, problems arise when the system of interactions becomes complicated so that the nonlinearities involved become many and involve causal loops. Indeed, the link from a wave process to quantal phenomena cannot occur without such nonlinear causal loops being present. For a complete detection/amplification process this is therefore always the case (see the lemmas set out above). Then the use of integration of differential elements becomes extremely heavy as a computation, and virtually useless as a means of visualisation and comprehension.

In the face of this difficulty we may investigate whether there are any other more economical ways of categorising, approximating and aggregating the classes of underlying waves such as to correspond to the total behaviour in the domain of observational results, other than by the inclusion of particle like entities. The most obvious of these alternatives is to stay with the original forms of expression in terms of spatio-temporal waves, and to improve the ways of describing these entities rather than the particulate ones. The two options are, after all, in a mathematical sense transmutable, at least in a crude sense and perhaps with precision. The prospect, and it turns out to be a promising one, is to find whether the transmutation back to the use purely of waves can be better than this crude sense ... to find whether this offers more convenient categories and structures for the modelling enterprise in these lower level processes associated with observation itself.

Coupled Observers

Low level couplings between the probes of multiple observers may be passive, but there is always some loss. As that loss approaches nothing so the coupling of the observations must become the dominant effect. For correlated observations the lowest possible level of loss is when two observers couple to a common system having little loss in either the couplings or the observed system (it is a nearly closed system except for the essential disturbances of the amplification). A case something like this is reported (refs [HSc00] [VGo]) at H.Schadwinkel et al on a University of Bonn web site, where a pair of avalanche photo diodes is brought to observe a single caesium atom in a magneto optical trap. However, the transmission paths of the couplings there encompass only some 4% of the total radiation from the atom in the relevant part of the spectrum. Thus the effects have to be evaluated in terms of marginal changes of correlation. Nevertheless the demonstration of a negative correlation between the output of the two detectors for short periods of time difference is interesting evidence. Those who think in terms of particles and instantaneous detection events call this "photon anti-bunching" and impute a non-local influence as being necessary to explain it. So long as we recognise the coupling in existence between the two detection processes and the significant time of detection resolution involving both detectors we need no such extension to the original causal model.

The special case of a pair of observers where each observer has as its object system only the probe of the other observer is interesting. Is there any observable correlation between the firing times of two closely coupled face-to-face avalanche photo diodes? This could be tested both with thermal dark observation and perhaps with a weakly coupled flow of light into one of the avalanche photo diodes (APDs). A reason why this might not show correlation could be that the regions of resolver activity in each diode may not be able to be sufficiently well selectively coupled (focused) onto corresponding resolver areas in the other. Without such explicit coupling each diode’s amplifier activity would be coupled to an average of the thermal activity of the other. Correlations would then possibly be overwhelmed by uncorrelated thermal noise, notwithstanding the detection process in each diode involving the other diode. Perhaps the experiment would need to be set up with a mode selector between the two detector diodes ... coupling via say a small aperture or fibre. A partial reflection of some sort could then serve as coupling for an external light source into that mode. In the case of the above mentioned MOT experiment at Bonn the atom itself acts as a mode selector. I am describing here something which looks like an "entanglement" experiment, but treating it only in terms of wave processes ... there is unlikely to be any need to invoke non-locality as an explanation of the correlations it produces, whether they be positive or negative.

Cybernetics Rationale for Quantisation.

See Gerhardt Grössing on circular causation [GGr00]. Though displaying a commendable adherence to circular causality as the basis for modelling, still the book by Grössing starts from a basis in the physics of discrete entities which I prefer to avoid, for the reasons explained above.

See Anton Zeilinger on the information basis for quantum mechanics. From Gary Boyd in the CYBCOM emailing list Tue, 20 Feb 2001 12:31:36 -0500 "Primacy of CyberneticsII for QM.?- [Zeilinger's] new work" quoting "In the beginning was the Bit", and the article by Hans Christian von Baeyer in New Scientist of 17 Feb 2001:
"Why does the world appear to be quantized?" Anton Zeilinger replies "Because information about the world is ((necessarily!? -GMB)) quantized". See also: Zeilinger's web site. Gary Boyd makes the point that "This seems to make cybernetics II the foundation of physics, rather than vice versa?"

Though I am sympathetic to Zeilinger's approach in which discreteness is accepted to be a necessary trait of a system of observation, my approach starts from a more general nature of "discrete" than just "binary", and seeks to show how such discreteness can emerge from a continuum. My approach is basically physical rather than information theoretic.

So I accept that the model of an observer must manifest discreteness in the nature of its recorded or propagated signals. However, I wish to show that the form of underlying physical model required need not, and for some purposes should best not, itself be conceived as built from discrete elements. Quantal effects would in general best be seen as emerging from the continuum as a concomitant of the inherent processes of observation. In other words any continuum which does not have structure sufficient to give rise to quantal entities can also not support any process which we would call "observation".

A programme of development might follow these steps:

Move beyond these into a Recursive Observation Lemma: Any general theory of observable reality that is valid for an observer must be valid in accounting for the processes in an observed observer.
Establish the logic that amplification would never be possible without some form of quantisation (see above).
Question where this quantisation is seated, and postulate, for room temperature observation, that it is in a system of attractor-like behaviour in electro-atomic structure.
Suggest that this can be woven into (if not simply derived from) the existing body of quantum electrodynamics.
Moot the possibility that similar reasoning might apply in quantum chromodynamics, but not to be developed here.

There are questions about the concept of attractors as means to the realisation of quantal behaviour. We may distinguish between two sorts of attractor mechanism which can be classified as:

1) AUTO-attractors: There is an ever present drift of state towards the attractors independent of the intensity of causal interaction. This covers also what is called hysteresis.
2) TRANS-attractors: Negligible rate of spontaneous drift occurs, but only evolution of state directed towards attractors at a rate depending upon both state and intensity of causal input activity.

It is the latter which is needed to account for the Einstein photo electric effect in which quantized effects occur with mean frequencies in proportion to a causal radiation influence down to very weak levels of radiation.

We can characterise these two mechanisms by way of distinct models derived from the Fokker-Planck type. The Fokker-Planck equation balances the effect of dispersion, usually seen as "diffusion", in slowly destroying a structure of space and time as by melting its form, against a so called "drift" term which acts to drive the function into some structured form. It is natural in the Fokker-Planck model for the strength of the drift term to follow the intensity of the driving activity, whereas the diffusion term does not necessarily do so. If the diffusion term has a form dependent solely upon state then we have the basis for an AUTO-attractor model. If the diffusion term contains very little spontaneous continuous part and is constituted mainly from chaotic forces resulting from the drift forcing activity then the model has the form of a TRANS-attractor system.

Since we may observe both the photo electric effect at the quantum level and hysteretic phenomena at a macroscopic level it is evidently possible to assemble the AUTO form from the TRANS form of attractors. To achieve this would seem to require in the model of the macroscopic process no more than the presence of some form of inherent agitation, such as zero point fluctuation. It appears that it is not possible to assemble TRANS-attractor from AUTO-attractor mechanisms except through a relatively elaborate structure comprising an ensemble of thermodynamic sub-processes. Is the concept of zero point fluctuation able to overcome this limitation too? I think not.

(Note: With trans-attractors there is what is for me the very interesting possibility of modelling or even designing time or rate independent adaptive/evolutionary/learning mechanism here too! Refs [ABo89] and [ABo95] relate.)

I did not want to get involved with it here, but to complete the cybernetics model we need a causal link from the ψ wave function via geometry to the potential mediating the ψ-dynamics. Such a link must perform the function of Maxwell's equations (vector waves) whilst maintaining Lorentz invariance.

Notes

The low level couplings are unitary, reciprocal in interactions, necessarily described in terms of what they conserve rather than what they develop. The amplifier cannot be so. Although some of the systems we (wish to) observe may be passive in the short term (compared with observer probe resolution time), when observed the overall structure cannot remain so … presence of the observer amplifier is essential.

By “passive in the short term” here is meant that the said system (sub-system) contains no process which resolves superposed quantum states in times short compared with the resolution time of the amplifier resolver which operates in the probe.

The conventional macroscopic meaning of the word “passive” does not have this restriction to “short term”. It relates to whether there is essentially a very large number of quantal processes in the observed sub-system between which a conservation rule operates. It permits those very commonly encountered cases where relatively fast thermally driven resolutions are occurring between processes over which an average can be taken by the detector probe and its resolver process.

References and Background Reading