The Journal of Philosophy, Science & Law

Manuscripts and Articles

Volume 3, March 2003

The Problem of Observation

Avner Levin*


* Faculty of Business at Ryerson University



There is perhaps no area more popular or in demand among legal academics and faculties of law recently than the amorphous area known alternatively as ‘law and science’ or ‘law and technology’ or ‘law, science and technology’ or some other combination of these terms. In truth, these areas of research are better described as ‘the law as applied to recent developments in science, technology, biology, etc.’ a description making for a less catchier title perhaps. Nevertheless, academics active in these areas research the law of new technology, or the law of scientific innovations, or the law of biological developments. They do not research law and science or law and technology or law and biology as such, i.e., as two disciplines sharing some common epistemology or ontology or enjoying other insightful analogies. Theirs is not an inquiry into the common foundations of science and law, but an investigation that has been carried out many times before, of the laws and legal doctrines relevant to specific forms of human behavior. In these instances, it is an investigation as to what rules and regulations should cover the newly developing frontiers ranging from computer networks to embryonic stem cells.

These are of course extremely important and worthwhile inquiries. In this article, however, I will not be concerned with this form of academic activity about the law. What will concern me here is that other way in which an inquiry into ‘law and science’ can be understood, as an investigation of how, and in what manner, law and science are similar. My argument, interestingly enough against the background of proliferation of such inquiries, is that law and science (I shall limit my discussion to the discipline of physics) share not only some interesting superficial analogies but also a deeper, common epistemological basis. More accurately, there are epistemological limits placed on both law and physics that are in all important respects similar. These limits on the ‘quality’ of knowledge imposed on both physicists and lawyers are known as a Problem of Observation existing in the disciplines of physics and law.

The term Problem of Observation is a term that was first coined by the Danish physicist and philosopher of science Niels Bohr.[1] It signified for Bohr, in a manner that I shall discuss below, the limits of attainable knowledge in physics as evidenced by theoretical developments and practical experiments in quantum mechanics at the time. It signified that the time-honored method of learning through theorization and experimentation, grouped together somewhat inaccurately in English by Bohr to be called Observation, suffered from a fundamental breakdown that imposed limits on what exactly could be learned and the manner in which it was learned, a breakdown that could be termed in what was a bit of an understatement as a ‘problem’. Hence, Bohr’s Problem of Observation.

Of course, what was for Bohr ‘merely’ a Problem of Observation signified for others, most famously Einstein, what could be called a ‘problem of reality’.[2] According to Bohr’s opponents the puzzles created by quantum physics indicated that the knowledge learned through quantum physics could not ultimately reflect reality. There had to be, according to Einstein, a more fundamental theory of physics that would, through the knowledge it provided, solve the puzzles of quantum physics. The debate between Einstein and Bohr was therefore a broad philosophical one. Einstein understood quantum physics to be an ontologically flawed theory. It made for an understanding of physics that was incomplete and inconsistent with other theories, such as Einstein’s theories of special and general relativity. Bohr thought that Quantum physics did not produce an inferior understanding of reality, but a philosophical realization that there are limits to the regular method of obtaining knowledge. According to Bohr, the special circumstances of quantum physics brought to the foreground an epistemological problem that potentially existed in many other disciplines. Although I will base my discussion of the Problem of Observation in law on Bohr’s philosophical insights into physics I should note, therefore, that Bohr’s insights were not accepted by a significant portion of physicists attempting to ‘make sense’ of quantum physics, and led at the time by none other than Einstein.

With that in mind, I will discuss Bohr’s Problem of Observation as it is manifested in physics, then distill from it along Bohr’s lines of inquiry its essential elements, those elements necessary (and put together, sufficient) for a Problem of Observation to appear in any discipline of academic research. After identifying these elements, I will show them to exist in the area of law and therefore conclude that a Problem of Observation exists in both physics and law despite these disciplines’ many obvious differences. Finally, I will indulge in a bit of speculation as to what the existence of a Problem of Observation in the study of law may bring about. Since academic legal activity has been conducting itself unaware of such an epistemological problem it will probably not come as a great surprise that a Problem of Observation means little to the great majority of work done in legal research (just as, coincidentally, it means little to the majority of work done in physics). Where the existence of a Problem of Observation does have an effect is in the area of jurisprudence and legal theory, which has been enjoying a renewed interest in methodology, and I shall discuss its implications on the renewed, ‘pragmatic’ way of conducting jurisprudence.[3]

One more caveat before I begin my discussion of physics and law. My argument will not be that legal reasoning (usually distinguished from or compared with moral reasoning)[4] and scientific reasoning are the same. That is, I do not argue that knowledge is gained or should be gained in law and its areas of interest in the same manner that knowledge is gained or is purported to be gained or should be gained in those disciplines of the exact sciences such as physics or mathematics. Although such an argument could be made, and indeed has been practiced by many (most notably perhaps by Austin)[5] it is not my argument here. Rather I will argue that the manner in which scientists interact with their subject matter, and legal academics with their subject matter, whatever these respective subject matters may be, is similar enough in a few crucial aspects to warrant the extrapolation of some epistemological conclusions from the philosophy of science into the philosophy of law.



It is a famous anecdote that by the turn of the nineteenth century physicists thought that save the minor problem known as ‘black body radiation’ all of the important problems in physics had been solved.[6] Little did they know that this ‘minor’ problem would prove such a can of worms to their aspirations that within a matter of decades the discipline of physics would be turned on its head, with strange and wonderful terms such as ‘uncertainty’ and ‘probability’ coming to the fore of scientific research. One of the physicists attempting to make sense of all these developments was Niels Bohr. The father of what today is known as the Copenhagen Interpretation of quantum mechanics, Bohr set out to develop the philosophy of science so that it would be able to furnish somewhat of an answer to the puzzles posed by the experiments of quantum mechanics, puzzles that seemed, disturbingly enough, to emanate from reality itself.

These ontological puzzles (or more accurately in Bohr’s understanding, apparently-ontological puzzles) ranged from the proper interpretation of light (was it composed of waves, or particles?) to the inability of physicists to measure certain attributes of particles that they were studying, attributes that seemed to be coupled so that an increased knowledge of one came at the expense of the other (the most famous pair is that of position and momentum). Unlike Einstein (and others) Bohr identified an epistemological (and not ontological) foundation (or lack thereof) common to all these puzzles, that he called a Problem of Observation. But before arriving at this conclusion Bohr took apart the elements of scientific research, as he understood them.

According to Bohr academic activity in physics was divisible into two major forms of ‘observation’.[7] One was the form of academic work known as theorization. Bohr called this form the mode of description.[8] The second was the form of academic work known as experimentation. Bohr called this mode the interactive mode of observation.[9] In order for academia to flourish (in physics at least) both interactive and descriptive (or non-interactive) modes of observation were not to impede each other. Once such an impediment existed a Problem of Observation occurred.

What these modes of observations consist of is pretty self-evident from their titles. The mode of theorization is quite simply the mode in which the physicist theorizes about whatever natural phenomenon the said physicist is currently preoccupied with, such as the nature of light, or gravity, or elementary particles, or multi-dimensional super-strings, whatever those may be. As an ideal mode of activity, it consists of the physicist detaching himself or herself from her subject matter, reality, and using the concepts of the discipline (such as the notions of particle, or wave, or energy, or matter) to construct a theory, i.e., an explanation utilizing these concepts to clarify whatever aspect was of interest to the physicist in the observed phenomenon.

It is important to note that this is an ideal description of scientific activity. It is not intended, nor did Bohr intend it, to convey how scientific inquiry actually transpires. What is of importance is that no matter how such research does transpire it is divisible along Bohr’s lines, into a mode of theorization and a mode of experimentation.[10]

The other mode of observation, the mode of experimentation or mode of interaction, is again, as its title conveys self-evident. In the mode of experimentation the physicist ventures into the laboratory, that scene of experiments, in order to conduct a series of active investigations into that aspect of the natural phenomenon the physicist is interested in. The purpose of the experiments is to expose or reveal a quantifiable observation about the nature of the phenomenon at hand. Such an experiment, if conducted according to the appropriate scientific methodology, is designed to yield objective, or at the very least inter-subjective results, i.e., results that can be repeated (and therefore verified) according to the same methodology by other researchers. So light may be cast through one or two slits on a screen beyond, and the emerging pattern recorded. Or the course of planets around stars may be recorded, or the trajectory of particles emerging from the collision of two particle beams in a particle accelerator, and so on.

Now it becomes clears from the description of these examples of experiments that experiments conducted without the aim of a certain theory in mind are few and far between. In other words, experiments are designed with certain theories in mind, and they serve (ideally) to support or refute these theories according to the results that they yield. Similarly, out of an experiment designed on the premises of a certain theory another theory may emerge as happened perhaps most famously in the cases of ‘black body radiation’ experiments and photoelectric cell experiments that ultimately led to the triumph of quantum mechanics over classical physics.

There is, therefore, a relationship or interaction between the two modes of observation as described by Bohr. The mode of theorization gives birth to the mode of experimentation, which in turn influences further theorization, which in turn prompts further experimentation, and so on and so forth. When something happens to this reciprocal relationship a Problem of Observation is created, and the two modes of observation can no longer support each other and the discipline they are conducted in. The alarming result is the seeming collapse of the acquirement of knowledge.[11]

Bohr identified two attributes of this reciprocal relationship between the two modes of observation that he perceived as necessary to prevent a Problem of Observation from occurring.[12] These attributes or elements can be discerned from an analysis of the methodological differences between ‘practicing’ the science of classical physics and ‘practicing’ the science of quantum mechanics.

Suppose I want to weigh a big, unwieldy, yet not too heavy textbook on tax laws and ways to evade them (hence the book’s size). I have scales on which I weigh myself every time I go on a diet but I cannot weigh my textbook on them since they are designed for people to stand on with both feet in order for them to function. My ingenious solution – I weigh myself with the book and then I weigh myself without the book. The difference between the two measurements is the weight of the book! My mastery of physics was enabled in this example by my assumption that I could extract myself, quite literally, from the equation of the book’s weight. My interaction with the experiment I was conducting, to use a more sophisticated terminology, was determinate. I was able to know (here is where my mastery of the laws of physics lies) in advance how my weight will affect the total weight of the book and myself, and therefore I was able to deduce, by conducting a series of experiments, the weight of the book. So the first attribute of classical physics which enables the relationship between theorization and experimentation is determinacy, the ability to know in advance and in a quantitative form the effect the physicist will have on the experiment conducted.

Suppose now that before conducting my weight experiments I was eating chocolate while leafing through my tax textbook. I accidentally smeared some pages with chocolate, and I am now concerned that I changed the weight of the book (I suppose I should be concerned about the effects of eating chocolate on my own weight, but only for dietary purposes, not for the purposes of the experiment since in both measurements the chocolate will already be inside of me.) For most purposes my concern would be unjustified. Why? Because the weight of the chocolate stain is quite negligible compared to weight of the book. I would be unable to detect any difference between the stained and pristine versions of my tax book using the scales I have at home. The second attribute of classical physics is therefore negligibility. It is safe to assume, in those experiments in which the physical physicist is not intended to be part and parcel of the experiment, that the affect the physicist will have on the experiment is negligible, and cannot significantly change its results. A car will continue to zoom at 100 MPH, to give another example, whether or not a policeman watches it. The officer’s vision (in this case his sight, or his radar, or his laser gun, are the ways that the officer interacts with the ‘experiment’ – the zooming car) will have no effect on the car’s velocity.

The two attributes a scientist possesses when interacting with her experiment in classical physics are therefore negligibility and determinacy. The affect of the scientist on the results of the experiment can always be said to be either negligible, or determinate, or both (but not necessarily both). Why are these two attributes important to Bohr and the Problem of Observation? They are important because their existence enables the reciprocal relationship between the two modes of observation as defined by Bohr. The classical physicist builds upon pre-existing notions in order to construct a theory of the natural phenomenon observed.[13] Out of the theory experiments are constructed, and conducted, in order to reaffirm or refute the theory in question. But it is only because the effects of the actual scientist upon the experiment are either negligible or determinate that one can truly deduce the impact the experiment has had upon theory. Were the effects not so negligible, or not so determinate, it would become extremely difficult for such deduction to take place. In fact, argues Bohr, a Problem of Observation would occur, and theory could no longer explicitly rely on experiment for its purposes (and vice versa). The reciprocal relationship would be broken.

That is exactly what happens, according to Bohr, in quantum mechanics. First, some experiments become inherently indeterminate. It is impossible to measure with the same accuracy both the position and the momentum of a particle, for example. One is always secured at the expense of the other according to the principles of quantum mechanics. The scientist measuring one parameter causes indeterminate changes in the measurement of the other. It is important to note that for Bohr these were truly epistemological limitations, neither technical measurement problems that will be solved one futuristic day, nor a reflection of ontological flaws in quantum physics.[14] Experiments in which interference patterns of light are created illustrate this point beautifully. According to the theory that understand lights as waves when one shines light through two slits an interference pattern (i.e., dark and light strips) is created on a screen behind the slits. This is a famous and a simple experiment, which has been conducted successfully many many times. According to the theory that understands light as particles, however light is constructed of, well, particles. It should be possible, according to this theory, to position a detector between each slit and the screen and measure where each particle of light passes on its way to the creation of the interference pattern. However, when such detectors are positioned no pattern emerges. It is as if the measurement of the particle-like attribute of light comes at the expense of its wave-like attribute and vice versa (no detectors yield a pattern of interference but no knowledge of where the particles of light went either). The insertion of detectors (that symbolize the scientist in this case) causes indeterminacy in the results of the experiment. It was meant to measure the wave-like attribute of light but due to the interaction of the scientist it can’t.

So one ‘fact of life’ of quantum mechanics is that in some cases the interaction of the scientist with the experiment leads to indeterminate results. Another ‘fact of life’ is that in some cases the interaction of the scientist with the experiment cannot be safely neglected. For instance, experiments that attempt to measure quantities like the energy or momentum of particles are interfered with by the very light used to measure the position of those particles. The light carries with it its own energy that is transferred to the particle which is the focus of inquiry, altering its ‘original’ energy and momentum as a result, so that it is impossible to determine what they were. The scales of energy carried by the light and by the particle are similar, hence it is impossible to neglect the affect the scientist has on the experiment. It is as if I smeared so much chocolate as to indeed alter the weight of my tax book. But negligibility (or lack thereof) is not only an attribute exhibited microscopically, although the microscopic is the boundary at which classical physics breaks down and quantum mechanics takes over. One of the more famous macroscopic examples of the effects of non-negligibility (and of indeterminacy) is the sad tale of Schrödinger’s cat.[15] In this rather brutal example a cat is trapped in a sealed box. An ampoule of cyanide gas is attached to the cage and a mechanism is constructed so that the particle released upon the decay of a certain radioactive element triggers the mechanism which then breaks the ampoule which then releases the gas which then kills the cat. All that is known about the radioactive element is that there is a 50% chance that it will decay within the hour. So what can the mad scientist conducting the experiment say about the status of the cat after one hour has passed? It is of course either dead, or alive. But scientifically the cat is ‘existing’ in a state that is a mixture of both outcomes, as long as the box remains shut. It is both dead and alive so to speak, although such a statement is of course, while mathematically correct, pure nonsense. The cat ‘becomes’ instantly dead, or alive, the moment it is observed by the scientist (say by opening the box). The interaction of the scientist with the experiment is in this case anything but negligible. It affects the results of the experiment and leads to an epistemological conclusion unobtainable by other means.

Note that if the interaction of the scientist with the experiment in all these examples was indeterminate but negligible (I don’t know exactly how much chocolate I smeared, but what does it matter anyway?) or significant (i.e., not negligible) but determinate (my weight is not negligible but I know what it is) one would not have a Problem of Observation, even when quantum mechanics are concerned. A Problem of Observation is created only when the interaction is both indeterminate and significant. In other words, while either attribute is necessary to the creation of a Problem of Observation, neither is sufficient.[16]

Now imagine that the interaction of scientist with experiment is both significant and indeterminate. What happens to the reciprocal relationship between experiment and theory? It breaks down. The scientist is unable to ascertain (or refute) the theory which led to the experiment (and the notions at its base) by means of the experiment’s results. And these results, in turn, cannot lead to the creation or modification of theory since they are tainted in a manner that, while significant, is undeterminable. The implication these results have on the notions and theories, which led to the experiment, which produced the results, is impossible to ascertain. Consider the (classical) notions of particle and wave that are at the basis of theories about the nature of law. These theories produce experiments, such as the two-slit interference experiment discussed above. The results of these experiments are then supposed to reinforce (or undermine) the theories about light and the notions at their base. But when the results vary with the introduction of detectors, (that theoretically are supposed to change nothing), i.e., with the interaction of the scientist with the experiment then they cannot be said to relate in any manner to the theories that prompted them. They neither reinforce nor refute them. The results of the experiment and the theory are estranged. They have nothing to do with each other. As a result, theorizing about light cannot progress. There can be no theoretical evolution, no realization which conceptualization of light is correct. Is it waves or is it particles? The experiments simply cannot help the scientist decide. Knowledge about light, in this manner, can no longer be gathered. There is a Problem of Observation.

How is the Problem of Observation solved, or worked around? After all, physicists have not abandoned their laboratories in droves after the theoretical development of quantum mechanics. One solution is to abandon the classical notions, and try and develop ‘modern’ notions. The talk about cats existing in a ‘mixture of life and death’, of particles’ ‘probability functions’ and ‘spins’, of light being both wave and particle, are all attempts to create new notions that will correspond better to mathematical formulations and gloss over the Problem of Observation. But Bohr (again, unlike others) thought these new notions do not really take the Problem of Observation away, but simply bury it in layers of jargon. Bohr’s first important insight was that the Problem of Observation was only a problem of observation, or investigation, or inquiry. It was, in other words, an epistemological problem and not an ontological problem. The puzzles and paradoxes of modern physics did not reflect a breakdown of reality but rather a breakdown or an acknowledgment of the limited manner in which scientists could ever hope to understand reality. In this Bohr resembled Einstein more than either would perhaps care to admit. Einstein too did not believe that the puzzles of quantum physics reflect reality, and devoted of course a significant amount of his time to perfecting ‘thought experiments’ in order to prove just that. Yet for Einstein the fact that quantum physics could not ‘decide’ between the notion of light as waves and the notion of light as particles, or the fact that quantum physics allowed one particle to affect another instantaneously meant that reality could not be as quantum physics portrayed it. For Bohr such puzzles reflected the inadequacy of our classically developed physical notions to understand a reality more complicated than these notions would allow for. And since Bohr also thought that classical notions were the only intuitive notions that could figure meaningfully in any theory,[17]i.e., in any attempt to simplify and explain interesting aspects of natural phenomena he arrived at the conclusion that the problem he was facing was epistemological, not ontological. The way could be found, and had to be found, therefore, to retain the classical notions that were at the foundations of theoretical knowledge and build further knowledge with them, despite the Problem of Observation. And this assumption was Bohr’s solution. Bohr simply postulated that classical notions that were the building blocks of the mode of theorization superceded any epistemological problem that might occur. Bohr decided that epistemologically the relationship between the two modes of observation was more fundamental than the Problem of Observation, and therefore had to be retained despite the Problem of Observation’s existence. After all, it did serve to promote knowledge in the many cases where no Problem of Observation arose. So Bohr came up with an idea known as Complementarity, which was namely the idea that the two modes of observation are more fundamental than the Problem of Observation. They always complement each other to form together the basis of knowledge in any given discipline, and they continue to complement each other in those instances where a Problem of Observation occurs.

The idea of Complementarity is somewhat of a letdown as a solution to the Problem of Observation. It is a workaround, a philosophical detour, constructed by Bohr around the pothole that is the Problem of Observation on the highway of knowledge. In the case of light, for example, it leads to the concession that scientists will never fully understand the nature of light. Light is dual in nature, and its complementary aspects, of particle and wave, are the only ways in which scientists can conceptualize and theorize about it. So some experiments will continue to enforce its wave-like nature, while others will continue to enforce its particle-like nature, but a decision will never be made between the two. Complementarity, or more accurately the Problem of Observation, is in this sense a realization of the limitations of the epistemology of physics, rather than an acknowledgement that quantum physics does not provide a complete and consistent understanding of reality.

Disappointments aside, the discussion of the epistemology of physics leads to the conclusion that in the exact science of physics a Problem of Observation occurs when two attributes coexist in the interaction of scientist with subject matter. These two attributes are the interaction’s significance, and the interaction’s indeterminacy. Details of quantum mechanics aside as well, these two attributes can conceivably and potentially occur whenever an attempt to advance knowledge is made, in any discipline. Indeed, his debate with Einstein whether the puzzles of quantum physics reflect its ontological shortcomings or a more general epistemological problem encouraged Bohr not to view the division of academic research into these two modes as limited to physics. In later work he attempted to extend it to disciplines such as biology and psychology, and others have sought to extend it to disciplines such as anthropology.[18] I argue below that despite the debate over it within the philosophy of science the Problem of Observation can be shown to exist, not potentially, and not conceivably, but specifically, in the study of law, but I am not even the first to attempt and apply Bohr’s work to law. Englard attempted to do just that in the area of private law, an attempt that was unsuccessful for several reasons which I shall not go into here, not the least of which was the failure, however, to recognize the shared epistemological basis of science and law, that I shall now demonstrate.[19]



Why is law like physics? The short answer of course is that law is nothing at all like physics. Law is not an exact science, there are no laboratories in which lawyers experiment, no mathematics with which legal academics construct their theories of law. So it seems a bit ridiculous, not to say pretentious, to attempt and draw an analogy between physics and law that would be meaningful and insightful to the study of law. Nevertheless I will show here that law and physics share the same epistemological problem, the Problem of Observation. In order to do so some form of analogy, even if only for the purposes of example and illustration, must be drawn between physics and law.

In Bohr’s ideal and perhaps naive discussion of the science of physics there was a scientist attempting to decipher the secrets of nature, there was of course nature itself (in form of whatever phenomenon or aspect of a phenomenon that was of interest), and there was the manner in which the scientist observed nature. This observation proceeded in two modes, the mode of theorization and the mode of experimentation, which could be considered, as an attempt at generalization, as a mode of interaction (experimentation) and a mode of non-interaction (theorization). Accordingly, here is my suggestion for the legal analogy: Bohr’s scientist is replaced with a participant in the legal system (a term that should be broadly understood as referring to anyone that discusses law or writes about law). Nature is replaced by law, and observation remains, in a sense, observation. So that instead of the scientist observing nature I now present you with the legal participant commenting on law.[20] So far, so simple, even if not very impressive or perhaps accurate. Of course, the foundation on which the Problem of Observation was able to grow was laid by Bohr with his suggestion that observation occurs in two modes. Accordingly, I propose two modes for legal writing. In one mode participants write about the law as it is, and in the other mode participants write about the law as it ought to be. I will therefore call these descriptive and prescriptive forms of legal writing.[21]

Now it is not my intention to classify every bit of legal theory and every last legal article as descriptive or prescriptive. In fact, although it could probably be argued that certain works are more descriptive or are more prescriptive I think that every work is a bit of mix of both and discusses the law as it is, the law as it should be, reasons for why the law is as it is, reasons for why the law should change, and so on. So I cannot stress enough how much I do not intend the above portrayal of legal writing to imply anything about how legal writing is actually done. In fact, there is probably smaller resemblance between my portrayal of the science of law and academic activity about the law than there is between Bohr’s portrayal of the science of physics and the work physicists actually do. What my division of legal writing into the two modes of description and prescription actually achieves is a much humbler goal. It enables the understanding of legal writing in a way conducive to the realization that a Problem of Observation exists in law.

It also helps me answer the following question: Why is law distinct from other disciplines? (That are also, presumably, eager to claim the Problem of Observation as their own as well.) The answer lies in the goals of prescriptive legal writing. For surely prescriptive legal writing is distinct in its aspirations from literary criticism or the weekly movie review. It brings forth not only its own interpretation of the law, but also concrete suggestions as to how the law should change. It is closer in kind therefore to the restaurant column more than it is to the movie column. For it deals with the law, work that is not static once completed (like a movie, or a book, or history, if history can be referred to as ‘work’)[22] but work that can change and does change in response to the needs and desires of society. The law is never a finished product, even if it is not necessarily “working itself pure”[23] according to some theoretical ideal.

Perhaps all legal activity, all that participants in the legal system are doing, can be understood as an attempt to create a consensus about how to bring about inherently controversial change, change that is laden with moral, social and political implications and that is ultimately, in some sense, binding. This tension in law is what sets it apart from other areas of inquiry in those disciplines known as the humanities or social sciences, and what distinguishes prescriptive legal work from criticism voiced within these other areas. It is also what distinguishes law as a discipline in which the Problem of Observation has the potential to emerge, as opposed to literature, or history, or sociology, or economics. There could be other analogies drawn between physics and these disciplines that might make even for compelling arguments as to other possible applications of Bohr’s Problem of Observation (anthropology is an obvious candidate, as I have mentioned above), but none share this unique feature of law.

Before I discuss when and where this Problem of Observation appears let me first illustrate how, in my portrayal of the discipline of law, observation, i.e., writing about the law, occurs. In order to do so I find it helpful to discuss the law on three artificial levels of generality, constructed solely for the purposes of clarification.[24] Those are, first, the level of particular norms, or rules, or laws.[25] Second, the level of legal doctrines, of areas of law, such as torts, contracts, aerospace law, and so on. Third, the level of law in general, where theories about law are offered from various perspectives, such as Hart’s, Dworkin’s, or Raz’s unique perspectives.[26]

Take a look at a particular law, such as that law often used in examples, a traffic law limiting vehicles to a certain speed on the highways. Some innovative research was recently published suggesting that if vehicles do not travel then road accidents do not occur. I seize the opportunity and write an article in which I argue that traffic should be brought to a stop, i.e., that the speed limit should be zero.[27] Much impressed by my argument and the support it has in empirical research the legislature convenes to amend the law and all traffic is brought to a halt. After a short period of living under these conditions someone else writes an article suggesting that perhaps traffic should be allowed to proceed at a certain speed after all, for reasons that have become a bit more obvious recently. The law is amended again, and so on, and so forth.

How does reciprocity happen in this example? The descriptive mode is the foundation upon which a prescriptive argument is constructed. This, in turn changes the law. The law’s description changes accordingly, prompting a different prescriptive argument, changing the law again. And again the law’s description changes, etc., etc. Note that in the example above the bit of writing actually describing the law was relatively trivial, as is sometimes (but certainly not always) the case when discussing particular laws. So it seems perhaps in the example above that descriptive legal writing is somewhat of an artificial construct. One needs only to think of slightly more complicated examples to realize that this is not the case. It is also important to realize that in my ideal example the articles I wrote, and the articles written presumably by my many outraged colleagues had their stated desired effect on the law. That is, the law changed exactly as prescriptive legal writing willed it to change. Unfortunately (or fortunately, as the prescription may go) that is not always the case. It is also the key to the realization that a Problem of Observation does exist in law. The vagaries of law’s normative change, as I shall discuss below, bring about this problem’s existence.[28]

It seems therefore pretty obvious that a reciprocal relationship exists between the two modes of legal writing, at least when focused on particular legal norms. What of the level of legal doctrine, or areas of law? Here too a reciprocal relationship exists although the discussion of law is both more general and more complex. Consider tort law. One participant might argue that tort law serves the purposes of distributive justice (i.e., restoring a pre-existing ratio between litigants according to their respective wealth, or wisdom, or height or whatever other criterion carries favor with the courts).[29] Another might argue that tort law serves the purposes of corrective justice (i.e., restoring or performing the compensatory equivalent of restoration of the state of affairs between litigants).[30] Both probably rely on what they attempt to be an objective description of tort law in support of what is, in effect, their prescriptive program. And both expect their prescriptive program to produce a corresponding change in tort law, in the adjudication of specific cases, and in the legislation of existing and future torts. This change will of course bring about a change in the description of tort law as relied upon by the various participants when detailing their prescriptive programs, and so the reciprocal relationship continues.

There does not exist therefore a fundamental distinction between the reciprocal relationship as it is carried out on the level of particular norms, and between the reciprocal relationship as it is carried out on the level of legal doctrine. What difference does exist relates to the discussion of legal doctrine being on the whole more complex than the discussion of specific laws or rulings, although that is not of course necessarily the case. There may therefore be a bit more description involved when discussing areas of law than when discussing particular laws in of themselves, and the call for change of areas of law is again possibly more complicated than a call for change of a specific ruling or piece of legislation would be. Be that as it may, the observation of law as it is conducted on both these levels remains basically unchanged.[31]

Finally, observations about law, the nature of law, the mechanism of law, the essence of law, and so on are generally given as well. These too can be divided into descriptive and prescriptive observations (although as I said I suspect every legal theory is a mix of both). Here too it can be said that descriptive observations form the basis for prescriptive observations, which then bring about a change in the description of law, leading to a change in the prescription of law and so on.

It seems such an undisputed fact that particular laws and areas of law are open to criticism and therefore open to change and revision.[32] It is perhaps a bit surprising that theories, notions of conceptions of law, are sometimes thought not to be so receptive to the possibility of change. This is a delicate point, and perhaps an example will do it some justice. One legal theory that is undoubtedly prescriptive in its aspirations (and therefore makes for an easy example) is Dworkin’s theory of law.[33] According to Dworkin law’s purpose is to justify political coercion in society, and such justification is best achieved when law enjoys a virtue Dworkin calls integrity.[34] As is the case with particular norms and areas of law Dworkin bases his analysis of law on a description of law in general, and several specific legal systems. According to this description the degree of integrity a legal system enjoys varies from system to system. Dworkin’s prescription, so to speak, is meant to enhance the integrity of systems that relatively lack it, and preserve it in those systems that enjoy it already. If judges act accordingly (Dworkin’s theory is directed mainly at judges, and has come under criticism for being a theory of adjudication rather than a theory of law)[35] then law will change and legal systems will change. The description upon which Dworkin rests his program for law will also change, and presumably Dworkin will finally rest his case, much in the manner that I did having convincing the legislature to abolish traffic.

Legal writing is therefore reciprocal in its relationship even when it comes to discussing law in general. Note that I have made no assumption about law’s openness to change. My only assumption is that legal writing necessarily presumes that law is open to change, at least when it is operating in its prescriptive mode.[36] Otherwise it has no point, and will indeed resemble literary criticism.[37] That the law of so many jurisdictions has changed time and time again proves this academic assumption is not so careless.

Now all these examples are of course a bit too easy. The legislature always does what the academics want, rulings and indeed legal reasoning change according to the dictates of legal theory, etc. Legal academics enjoy, according to this pretty picture, unrivalled success and influence. The sad truth is that this is not the case. However, it is a happy fact for my purposes, since it demonstrates that the conditions necessary for a Problem of Observation to occur in law do exist. My claim, therefore, is that the actual circumstances of the interaction of participants with law cause a Problem of Observation to appear in law, and disrupt the reciprocal relationship between the mode of description and the mode of prescription. These circumstances are the direct result of conflicting claims about law’s purpose (or purposes) on whatever level of generality one chooses to discuss the law.

The real circumstances of legal writing are that one never knows the normative effects a specific piece of writing will have upon the law. Change in law is subject to so many conflicting political and social interests, not to mention straightforward normative arguments, that the effects of one specific writing are impossible to predict in advance.[38] At the same time, this matrix of interests and arguments makes it impossible to decide in advance that particular works could be safely ignored. The arguments of one article may suddenly fit far-flung interests in their end results so as to elevate it into favor and cause it to have great influence over the legal issue discussed.[39]

Yet if these two aspects of the sad reality of the participant’s life seem familiar it is because they echo exactly those two attributes of quantum mechanics that cause a Problem of Observation to occur in physics. Those two were the indeterminacy of the interaction and its untold significance (i.e., the impossibility to safely neglect it). What is the indeterminacy of the legal interaction if not the inability to determine the effects writing will have upon the law? And what is its untold significance if not exactly that, the inability to discard with certainty certain writing as irrelevant?[40] The interaction of the participant with the law (i.e., the prescriptive mode of observation) is therefore done under two constraints: First, the change it causes in the law is indeterminate in advance, and second, the change it causes in the law can never safely be neglected in advance. Since change in the law is essentially unpredictable the reciprocal relationship between descriptive legal writing and prescriptive legal writing is broken, and a Problem of Observation occurs.

It so happens that the undisputed majority of participants writing about the law are blissfully unaware that the study of law suffers from a Problem of Observation. But this is not as surprising or detrimental to my argument as it may seem. A lot of work is done in physics and other sciences without the Problem of Observation rearing its ugly head as well, even though in principle it continues to exist. That is because in many cases it is possible to safely assume that certain works will be neglected, or it is possible to come up with a sufficient approximation of the effects a certain article will have upon the law. For instance, it is probably safe to assume that articles about the law written by tenured professors at Ivy League institutions will change the law more than articles written by students at non-credited law schools. The key point here is that it is safe to assume this in a lot of cases and circumstances, but not in all of them.[41] So the reciprocal relationship between the law’s description and its prescription continues to peacefully exist most of the time, just as it does in physics between theory and experiment, and perhaps in other disciplines as well. The Problem of Observation certainly does not permeate every aspect of legal life.

Make no mistake about its existence, however. An illuminating and convincing argument as to its existence lies by way of counter-example. Suppose that what holds true in most circumstances is true in every circumstance. That is, students from certain schools are never treated seriously and Ivy League professors are always treated seriously and so on. If this were indeed the case then we would probably have by now some form of ‘conclusive-argument-clincher’ grading that we could apply to legal writing. It would state, for example that a student’s work is always negligible, a graduate student’s work negligible if it pertains to public law, a non-tenured professor’s work negligible if it is about constitutional law or something similarly silly along these lines. Such a form of ‘negligibility-grading-in-advance’ seems to me absurd. It is absurd because the legal world would like to believe that its writing will be judged according to its merit (in terms here not of overall academic quality, although that too is an important consideration, but of its efficacy at securing normative change) and not according to its pedigree (i.e., one’s institutional affiliation). Of course pedigree is one popular rule of thumb as to the existence of merit, but the correlation between pedigree and merit does not necessarily exist in each and every case.[42]

The same argument can be constructed with respect to our ability to know in advance how particular writing will change the law. Here too, if this were the case then one would be able to predict normative changes as a routine manner. But it so happens that arguments that appear very convincing when given from a certain perspective lose favor when other considerations are taken into account, and their affect on the law is therefore mitigated. Similarly, arguments with little to them may suddenly gain favor in the eyes of the courts or legislature and bring about greater change in the law that could have been anticipated. The argument that tort law should rest solely on corrective considerations may appear to some to be more philosophically sound than the opposing argument basing tort law on distributive considerations as well. So it could be thought in advance that it would bring about change in the law as adjudicators, legislators, lawyers and other participants recognize its validity. However, the shear magnitude of social and political factors urging tort law to take distributive concerns into consideration as well may cause the argument basing tort law on corrective justice to be overwhelmed, rejected, and largely ignored. Of course the opposite might hold as well. One never knows in advance what normative change a particular argument or prescriptive observation will bring to the law.

The unrealistic nature of these counter-examples supports, therefore the conclusion that no determination in advance can be made with respect to prescriptions about the law, in its two aspects relevant to my discussion. Prescriptions about the law cannot be unceremoniously ignored, and the actual change they will bring about in the law cannot be determined in advance. Although neither of these aspects is sufficient, on its own, to create a Problem of Observation in the study of law both are necessary to its conception and their mutual (continuous – this is not a redundant requirement in law as I shall shortly discuss) existence is a sufficient condition for a Problem of Observation to exist.

As ludicrous as it may have seemed at first, it appears that physics and law are epistemologically similar, at least as far as the pursuit of knowledge is concerned. In both disciplines the participant is interested in increasing knowledge, so in both disciplines theories about their subject matter are constructed. In both disciplines these theories are then measured against a state of affairs that does not directly rely upon these theories for its ontological foundation. In both disciplines this measurement of theory, or interaction with the current state of affairs brings about an indiscernible change in the very state of affairs the theory is set to address. And in both disciplines the understanding of their subject matter is disturbed as a result of this change, the participant cannot uninterruptedly rely on this understanding in order to improve upon it, and knowledge cannot be freely increased. Both disciplines suffer from a Problem of Observation.

I need not remind the reader that this epistemological similarity does not equate physics and law. In physics, after all, Bohr argued that the Problem of Observation reflects an epistemological, rather than ontological issue.[43] In other words, scientists pursue knowledge in physics without the intention (and Bohr argued, ability) of changing the state of nature. The uncertain quality of their measurements detracts from their ability to theorize about nature so that what seems to them an ontological fact (for instance, that light is both particle and wave-like in its essence) is in fact a reflection of their epistemological limitations. In law participants set out with the intent of changing the law, their subject matter. If, in my crude analogy between physics and law the ontology of law is made up of the laws of society then one crucial, and of course well-known, difference between physics and law is that in physics observation, the tool of epistemology, cannot change what is, whereas in law observation as I have characterized it, although still epistemology’s tool, works to change what is law according to what ought to be law on a regular basis. So the Problem of Observation in law is not manifested, as it is in physics, through a warped perception of (legal) reality. The Problem of Observation in law is, in this sense, much less of a problem. It does not reflect the fundamental difficulty physicists have to perceive the laws of physics as they truly are. There is no such difficulty for academics studying the law. Or so it would seem. In the next section I will examine closely this assumption, especially as it pertains to that third level of generality when discussing the law, the level of theories about the law in general.



What is the significance of the Problem of Observation for legal writing? In order to answer this question I must first discuss its apparent insignificance for legal writing about particular laws and (perhaps to a lesser degree) about doctrinal areas of law. The recognition of this insignificance emanates from the realization that the idea of the Problem of Observation is meant to provide a partial answer to an (according to Bohr, merely apparent) ontological problem. So when this apparent ontological problem is termed a Problem of Observation insight is gained that whatever difficulties research in the discipline at hand seems to suffer from, they are epistemological difficulties, not ontological ones. But as far as law, or let me discuss initially, particular laws are concerned, it is hard to conceive of even an apparent ontological difficulty. There are no puzzles or paradoxes that appear to confound the legal observer about particular laws (let us leave aside for the moment objections to legislation and adjudication on political, moral and social grounds). In fact, legal reality, or legal practice (to the extent that these two terms are synonymous and at all meaningful) appear to be rather ordinary. Accordingly, writing about particular laws appears to go about its non-Problem of Observation course, so that despite the inherent significance and indeterminacy of prescriptive writing (which necessarily exists as I have discussed above) prescription is able to lead to change in the law, to a new description of the law, to an alternate prescription for the law, and so on.

Note that the circumstances here are different than the circumstances discussed in the previous section. There I mentioned those instances whether the interaction between participant and law can be safely neglected or determined in advance, and therefore a Problem of Observation does not arise, even though it is in principle present (mirroring physics, where classical assumptions can be made in certain circumstances, even though the underlying theoretical foundation is necessarily that of quantum mechanics). Here my assumption is that we are already under the conditions in which a Problem of Observation could exist. That is, my assumption is that the effects of the particular prescriptive writing cannot be summarily dismissed or readily discerned in advance. Still, it is difficult to conceive of an example (since I have not yet discussed the importance of the continuity of interference with observation in law) in which some question about the existence of a particular law could arise, and then be partially answered by the recognition that it is an epistemological rather than ontological problem. At the same time, however, it is commonly observed that ‘the law is silent’ on certain issues, or that there is no law regarding certain matters, or that no one knows what area of law is applicable, or that the nature of law itself is controversial.[44] I will argue that such observations about law, whether they are about law in general, or about areas of law or particular rules, observations that appear to be ontological observations about the law, are examples of those instances in which a Problem of Observation in law does occur.

Imagine if you will the following legal triangle. At two corners stands the participant, split for the sake of the argument into a descriptive and prescriptive version. At the third corner stands reality. The lines of argument flow thus: Description from the descriptive participant to the prescriptive participant, prescription from the prescriptive participant to reality, and change from reality to the descriptive participant. A new cycle can then begin. (Its starting point is of course arbitrary.) As long as there is no apparent ontological problem reality exists secure in its corner, and the lines of observation flow freely, and in sequence. But should such a problem appear to exist then reality would vanish, and the triangle would collapse.

Now imagine a triangle of physics. At two corners stand the theoretician and experimenter, and reality stands at the third corner, cornered in by the metaphor so to speak. Theory flows from theoretician to experimenter, experiment from experimenter to reality, and results from reality to theoretician. When the results are unobtainable it appears as if reality is unobtainable and an ontological problem has occurred, but it is actually ‘only’ a Problem of Observation.

The difference between the two triangles is that whereas in physics the results are unobtainable, and appear as a change in reality, in law change is readily observable after it has already happened (as far as particular laws are concerned at least). There is no problem discerning how the law has changed after the fact, even if before it was impossible to predict with precision the effect a particular prescriptive argument would have upon a particular law. There is no continuity of interference. Since the state of particular laws after each and every prescription is readily obtainable an ontological problem does not seem to appear, and needless to say, solving it by means of declaring it an epistemological problem is unnecessary as well.

It is clear, therefore, (to the extent that this analogy between physics and law clarifies anything about the law) that an ontological problem would seem to appear in law only when the results of prescriptive legal writing would be, in some sense, unobtainable, leading to the mistaken assumption that it is impossible for legal writing to decipher the nature of law. As far as particular laws are concerned such an occurrence is difficult to imagine, and in order to imagine such an occurrence I must further develop my crude analogy between physics and law, to determine in what manner interference with observation in law can continue, so that the results of prescriptive legal writing would become unobtainable.

Whereas such terms as objectivity, reality and the rules of nature appear to be well-defined and understood in physics and other exact sciences (at least for the moment) it is obvious that their legal counterparts fall under far greater scrutiny and disagreement. A closer examination of the terms equivalent within legal discourse (if such a notion exists) to these scientific terms draws one inevitably into the quagmire that is jurisprudence. Jurisprudence has moved over the years from the assumption (or presumption) that law can be understood by studying it from afar, noting its common features and analyzing them,[45] to the realization that the reality of law cannot be observed like the reality of physics, mainly because the reality of law, unlike the reality of physics includes those members of society on whom the law operates, its participants. The reality of law, in other words, is composed not only of bare legislation and adjudication but also of society, for which these laws exist. To ignore the manner in which society perceives the law would therefore be to miss out on an important illuminating perspective about the law.[46]

The realization (which in hindsight appears pretty obvious) that the reality of law is distinct from the reality of physics, and that therefore the methodology of legal theoretical inquiry should be distinct from the methodology of scientific inquiry leads of course to the realization that what appear to be ontological problems in physics will probably emerge under different conditions then what appear to be ontological problems in law. Indeed, it partially explains why it is that there are less of these apparently-ontological problems in law to begin with. Clearly these legal conditions will have something to do with the insight that society is part of legal reality. But although philosophers of law largely agree that society plays a role in their quest for understanding law, they largely disagree as to what that role actually is. The perception of law, of what law is, the understanding of law by participants, plays a role in accounting for what law actually is (to the extent that these are two distinct notions) that differs from theory to theory.

Clearly some philosophers of law would disagree.[47] The law is what their particular theories hold to be, regardless of how members of society perceive it. It may be, these theorists would say, that some theories hold such a perception to be of great importance, but that does not mean that their contribution towards the improved understanding of law is in any way dependent on the manner in which people actually think about the law. Ironically, these theorists think about law a lot like physicists think about the laws of nature. Law is ‘out there’, its nature or essence or some other important aspect of it waiting to be discerned and distilled by the perceptive theoretician, regardless and in many cases despite what popular opinion might take it to be. But law is not like physics. It changes whereas physical reality, remember, at most appears to change. It changes not as a result of a force of (legal) nature but as a result of the constant interaction members of society, participants, have with their law. Its individual norms reflect society’s particular values (with great or little success, as the case may be) and so do its doctrinal areas. Why suddenly assume, therefore, that the understanding of law in its entirety, the understanding of law’s essential features (that change of course from theory to theory) is not influenced by the same interaction?

And of course, not all legal theorists hold that the notion of law can be distilled without reference to its popular perception. That legal theory should rely on the popular perception of law is one crude way of summarizing ‘pragmatic’ jurisprudence.[48] Indeed, one simplistic distinction between those two big tents of jurisprudence known as positivism and anti-positivism is that positivism studies the law as it is, whereas anti-positivism studies the law as it should be. Although simplistic and inaccurate this distinction nonetheless captures the emphasis some theorists put on equating the theory of law with law itself, while other theorists perceive a theory of law as a goal for law to achieve. What is of importance for my purposes, however, is the recognition that a debate exists within jurisprudence as to the role the perception of law by its participants plays in the construction of theories of law. That such a debate is even able to meaningfully exist within jurisprudence indicates how distinct legal reality is from physical reality. The existence of this debate justifies, therefore, the following crucial ‘pragmatic’ (or rather, empirical) modification of my analogy between physics and law: Whereas in physics the notion of reality is independent of the opinions physicists hold with respect to nature, in law the notion of law is dependent on how the law is theoretically perceived by participants, legal academics (and their rejection of this assumption within the theories they put forward) notwithstanding. And now I can ask the following: What is the significance of the conclusion that the notion of law is dependent on its perception by members of society? In what (if at all) circumstances does it bring about an apparently-ontological problem in law?

Since an apparently-ontological problem is one in which it there seems to be a problem with legal reality, and since legal reality is constructed, in part, on the basis of the perceptions members of society have about the law then it appears that one instance in which an apparently-ontological problem will come about is when it is impossible to determine in advance or safely neglect the effects a particular prescription will have, not only on law in general, but on these perceptions in particular. When an interaction between a prescription for law and the perception of law occurs, when a prescription for law seeks to change legal reality by directing itself at the perceptions of law, then a continuous interference with the observation of law occurs, and it is then an apparently-ontological problem will occur. It would become impossible for philosophers of law that seek to describe the law, and explain it through its description, to rely on legal reality, at least in its part that relies on the perceptions of members of society, in support of their explanation. Any analysis of law, they would therefore conclude, must be independent of what the law is perceived to be. That some legal philosophers are in fact arguing strongly for this position is an indication, therefore, that they are faced with what seems to them to be an ontological problem.

But on what level of generality, so to speak, is a prescription directed at the perception of law? The short answer is that it is directed at the perception of law at all levels. My distinction between particular rules, areas of law, and law in general is, recall, artificial. It is of no substance when discussing the perception of law, since the perception of law involves all of these levels. When a particular prescription advocates change in a particular law questions of the interpretation of particular norms, their limits, their sources, their precedents, and so on arise, and these are all inescapably connected to the perception of the area of law with which the particular norm is connected, and to the perception of law in general. These perceptions cannot, in practice, be disentangled into separate and distinct perceptions. One is then left with a case-by-case analysis. Sometimes there is such a consensus surrounding the particular norm in question, such as my traffic law, that the prescription aimed at it hardly bothers more general perceptions of law. Sometimes the norm in question is highly controversial, such as the norms of abortion, or capital punishment. People then say they do not know what the law is regarding that particular issue, emphasizing not their lack of knowledge, but what they perceive as a lack of law. But this is merely a restatement of the apparently-ontological problem that appears to them. Participants truly believe that there is a problem with legal reality when such issues are discussed, that there is no law in existence. My discussion has aimed to show that such a conclusion is mistaken. The law does exist, and the problem that appears to be ontological is in effect epistemological, a problem of not knowing what the law is, a Problem of Observation.[49]

Areas of law are subject to case-by-case analysis as well. The perceptions of what these areas are (for instance, what tort law is about, or what privacy law is about, and so on), are shaped on the one hand by the particular laws that constitute them and how these individual norms are perceived. On the other hand, prescriptive work is at times directed at areas of law in general and specifically at the popular perceptions of these areas, seeking to change them, and through them, that particular area of law. So apparently-ontological problems can arise, in these circumstances, in two forms. There can be a dispute over the nature of an area of law, which reflects contradicting perceptions about that area. For example, there can be a dispute whether tort law incorporates distributive concerns or corrective concerns. Problems of this type are usually formulated as problems over what the nature or essence or idea of the area of law is. Another form of apparently-ontological problems focuses on controversial issues and their allocation to areas of law. How should an agreement between a couple and a woman, that she hand over her child to them be judged, for instance? Is it a matter for contract law? Is it a matter for constitutional law, a question of human rights? Questions such as these signify that the reach of an area of law, its boundaries and scope, are in doubt. This is also, therefore, an apparently-ontological problem. In both cases the problem is created when the prescription affects the perception of the area of law in an indeterminate yet significant manner so that in both cases it seems that there is a problem with the area of law. In both cases, however, the problem is, again, epistemological and not ontological, a Problem of Observation.

Jurisprudence always aims to affect general perceptions of law. Perceptions of law in general are affected less by prescriptions of particular laws (except in exceptional or outrageous instances that have the potential to change the perception of law in general), and less by prescriptions directed at understanding areas of law differently, than they are affected by prescriptive work addressing law on its most general level directly.[50] It is in jurisprudence that the greatest controversies surrounding the nature of law arise, precisely for this reason. Since there are constant arguments in jurisprudence over the law philosophers of law reasonably assume that the nature of law is in doubt. They routinely ask ‘what is law?’, an apparently-ontological problem. Yet here too an epistemological problem is in fact at play, the same Problem of Observation. Let me return again to my legal triangle in order to illustrate this point. When a philosopher of law puts forward a theory of law she is, to use my terms, offering a prescription for how law should be perceived, and therefore offering a prescription for how the law should change. But since it is impossible to predict in advance how perceptions about the law will change, or even to suggest in advance that the philosopher’s theory about law could be neglected, any attempt to later describe the law cannot succeed, since it depends on the very perceptions the theory of law was directed at and these, in turn, are subject to continuous interference. Members of society presumably become aware of the theory of law, yet the manner in which their perception of law changes remains undefined. Some participants will adopt a new perception, some will remain unchanged in their own convictions, some will seek a perceptional middle ground, and so on. The total, or representative, or common, or consensual perception of law will remain, however, vague.[51] Since any later description of law relies on obtaining information about legal reality, and since the perception of law by members of society is in effect, unobtainable, the outcome is that academic work that purports to be descriptive (emanating from that other corner of the triangle), i.e., that purports to tell us what law is, has what are actually prescriptive implications. Rather than merely accounting for legal reality (which it cannot do in these circumstances) it has unintended effects on the perception of law by members of society.[52]

Perhaps a concrete example is in order. Consider Raz’s theory of law. According to Raz the law is a social institution that claims authority over members of society.[53] It mediates between members of society and their reasons for action, but in order to be able to mediate between members of society and their reasons for action the law itself must not rely on these very reasons as its source, even partially. It must not refer members of society back to those very reasons it purports to replace as an authority. Otherwise it would not be able to claim authority and fulfil its social role as Raz understands it.[54] That is why Raz is considered an exclusive positivist, in contrast with those (inclusive) positivists that argue that the law is able to rely on these reasons, at least partially.[55] So the question in dispute (or actually, one question in dispute) between these schools of positivism is how to make sense of certain legal norms (whether legislation or precedent) in certain jurisdictions that appear to rely, or refer to reasons for action, such as morality. Inclusive positivists argue that these are part and parcel of the law.[56] Exclusive positivists argue that these are best understood as directives of adjudication, since according to them there can be extra-legal principles of adjudication (Dworkin, of course, who dismisses all forms of positivism, would disagree.)[57]

Now this debate between exclusive and inclusive positivists is an excellent example of how the Problem of Observation confounds jurisprudence, once it is considered according to the terminology I have developed above. Raz sets out initially, with his understanding of law as an institution that claims authority, to offer what is primarily a description of law. He sets out to tell us what law is, not how the law should be. Confronted, however with a debate that addresses the perceptions of participants directly (how are these legal norms that refer to morality be understood, and is there any argumentative force to their popular perception as law?)[58] Raz’s insistence that these directives are best understood as directives of adjudication carries with it an unintended prescriptive consequence. He is understood as arguing (even if this is not his intention) that these directives should be understood (by participants in the legal system, legal theorists notwithstanding) as directives of adjudication. So work that sets out to be descriptive turns out to have unintended prescriptive implications, as a result of an ongoing ambiguity within the perception of participants.

The result of examples such as this is that for philosophers of law the nature of law seems controversial, unclear and subject to debate. There appears to be a problem with it since no one can state what it actually is, and therefore there appears to be an ontological problem. But if anything at all can be learned from physics in law it is this lesson: In this case what appears to be an ontological problem is an epistemological problem. Law is not in of itself problematic, but obtaining knowledge about the law is. Knowledge about the law is unobtainable in the circumstances I have just described but it is important to recognize that this is a problem emanating from the manner in which legal philosophers (and philosophers of science) obtain their knowledge, and not an inherent problem of law. It is a problem of the observation of law which necessarily includes law’s participants’ perceptions, and not of law itself.


Solving the Problem

The realization that the problems he was facing in the philosophy of science were epistemological rather than ontological was only the first step for Bohr. The Problem of Observation had to in turn be solved, or sidestepped, even once it was recognized to apply to knowledge rather than reality. What are the ramifications, therefore, of arguing that disputes commonly thought to be about what particular laws are, or what areas of law cover, or of what the nature of law is, are in fact not disputes over these issues but rather manifestations of the inherent difficulty in obtaining knowledge about the law since it is partially constituted by the manner in which members of society perceive it? What are the implications of the Problem of Observation for legal theory and legal philosophy?

Since disputes over particular laws and areas of law are so intertwined with normative, social and political arguments over the desired legal result it seems pointless to argue that the Problem of Observation will add any value to these debates. Any consequences it may have will quite simply be overwhelmed by these far more vivid and relevant arguments, and rightly so, for these controversies are not, ultimately, about law’s ontology and epistemology and they should not be ended by any conclusions that the Problem of Observation may lead to. Of course, jurisprudence too is subject to moral and political arguments, and those who believe that one should become a positivist (or anti-positivist) because of positivism’s (or anti-positivism’s) moral value are likely to be unmoved as well by any significance the Problem of Observation may hold to this or that particular legal theory.[59] Still, there might be a lesson here for legal philosophy to learn from Bohr’s philosophy of science, that might at the very least allow philosophers of law and participants in general to understand jurisprudence from a new and illuminating perspective.

That lesson is this: Recall that in physics, Bohr’s next step was to ‘solve’ the Problem of Observation by concluding that since it was an epistemological and not an ontological problem it had to be overridden by insisting that the two modes of observation were fundamental modes of epistemology. It was only through them that knowledge was gained. The fact that a Problem of Observation appears should not lead physicists to the conclusion that their modes of observation have to be discarded as inadequate. It should rather lead them to the realization that they have no choice and are forced to continue their search for knowledge using concepts and notions and ideas that cannot be completely supported by empirical data. Those notions are their building blocks of physical knowledge, and cannot be replaced or rejected. Therefore, the two modes of observation had to be retained no matter what the circumstances were, or what Problems of Observation arose. They continued to complement each other despite the apparent ontological, and actual epistemological, breakdown. Accordingly, Bohr called his solution to the Problem of Observation the idea of Complementarity. Although Complementarity is a bit of a philosophical anticlimax it nevertheless holds insight for law, and jurisprudence in particular.

Translating Complementarity from the language of physics into law it becomes apparent that the lesson learnt in law as well is that the two modes of legal observation are more fundamental than any apparent epistemological problem that the observation of law suffers from. So in jurisprudence in particular it becomes apparent that descriptive work and prescriptive work both continue to be necessary to the advancement of knowledge about the law despite the ongoing debates within jurisprudence about the nature and reality of law, and the necessity (or not) of developing a ‘pragmatic’ jurisprudence. These debates render a description of law unobtainable and tainted by prescriptive considerations according to my discussion. Furthermore, the idea of complementarity shows that debates over which conception or idea or theory of law is ‘correct’ or better fits the reality of law are especially pointless, despite their superficial appeal. Theories that are predominately descriptive (as some would argue Hart’s theory of law is) and theories that are predominately prescriptive (as some would argue Dworkin’s theory of law is) will all continue to be necessary building blocks of the knowledge of law for as long as there is a substantial lack of unity where the perceptions of participants are concerned. There is no point to ‘choosing’ between them, since they reflect the fundamental manner in which knowledge about law is obtained.[60] So long as we understand law as being a product of our perceptions, this is the unfortunate reality of our knowledge about the law.


[1] See N. Bohr, Essays 1958-1962 on Atomic Physics and Human Knowledge (1963) as well as his “Discussion with Einstein” in P. Schilpp (ed.) Albert Einstein: Philosopher-Scientist (1949)

[2] See P. Schilpp (ed.) Albert Einstein: Philosopher-Scientist (1949).

[3] See for instance J. Coleman, The Practice of Principle (2001).

[4] See for example J. Raz, “The Purity of the Pure Theory” in R. Tur and W. Twining (eds.) Essays on Kelsen (1985).

[5] J. Austin, Lectures on Jurisprudence (5th ed. Rev. 1885).

[6] A ‘black body’ is a body that completely absorbs all the energy directed at it. It heats up as a result and emits energy in turn. When physicists attempted to calculate this radiation on the (classical) assumption that the energy spectrum is continuous they received an infinite result, a contradiction with experimental data and an embarrassment. It was only when they assumed that the energy spectrum was discrete (hence quantum physics) that their calculations fit reality.

[7] Again, the English terminology is slightly misleading. Better terms are ‘investigation’, ‘inquiry’, or ‘research’.

[8] N. Bohr, Essays 1958-1962 on Atomic Physics and Human Knowledge (1963) at 6.

[9] Id. at 4.

[10] That is why Bohr’s conclusions are unaffected by the important work done on the actuality of scientific research from Popper to Kuhn. See K. Popper, Conjectures and Refutations (1963); T. Kuhn, The Structure of Scientific Revolutions (2 ed. 1970).

[11] Since that does not happen often there is apparently a way to work around the Problem of Observation, which I shall discuss below.

[12] I should note that Bohr’s ideas on the Problem of Observation evolved greatly throughout the years, and I am presenting them here in their final form. Initially, for example, Bohr assumed that the Problem of Observation emanated from difficulties internal to the mode of experimentation. For more on how Bohr developed his ideas about epistemology see H. Folse, The Philosophy of Niels Bohr (1985) and D. Murdoch, Niels Bohr’s Philosophy of Physics(1987).

[13] In truth they too are the products of the reciprocal relationship between theorization and experimentation, so this is an arbitrary starting point for the argument I am about to make.

[14] As I discuss below, Bohr viewed classical notions (such as position and momentum) as indispensable ontological building blocks of any theory of physics, including quantum physics. From this starting point quantum physics’ inability to obtain knowledge of both these notions simultaneously can only reflect an epistemological, and not ontological, problem, while the ability to conduct separate measurements within quantum physics of either of these notions neither strengthens their ontological ‘status’ nor eliminates the epistemological problem from which the theory suffers.

[15] For the English translation see J. Trimmer, The Present Situation in Quantum Mechanics: A Translation of Schrödinger’s “Cat Paradox” Paper 124 Proceedings of the American Philosophical Society 323 (1980).

[16] Together they are sufficient, but only if they ‘continue’ to interfere with the observation. This is a redundant requirement in physics, but it will be significant to my discussion of law.

[17] N. Bohr, “Discussion with Einstein” in P. Schilpp (ed.) Albert Einstein: Philosopher-Scientist (1949) at 209-210.

[18] Id. at 224, 236 as well as supra note 8 at 20-22; Folse, supra note 12 at 173-174.

[19] See A. Levin, Quantum Physics in Private Law, 14 CJLJ 249 (2001).

[20] Writing about the law does not imply that anything mysterious such as legal reasoning is going on. So when I refer below to legal writing it is, again, a simple term with no strings attached, referring to writing produced about the law by participants that include, among others, academics, judges, and lawyers.

[21] It may be quite tempting to identify positive theories of law with a descriptive approach, and natural-law theories, or anti-positive theories of law with the prescriptive approach, but one should resist the temptation. For example, is Raz’s insistence that the law claims to be an authority mediating between participants and their reasons for action descriptive, or prescriptive?

[22] All are of course subject to interpretation that brings out various and perhaps contrasting aspects, but not to interpretation that seeks to bring about their change.

[23] Omychund v. Barker 26 Eng. Rep. 15, 23 (1744).

[24] They are particularly artificial since participants perceive them as connected, in a manner that will become clearer below.

[25] I will assume these terms are all interchangeable for my purposes here.

[26] H.L.A. Hart, The Concept of Law (P. Bulloch and J. Raz eds. 2nd ed. 1994); R. Dworkin, Law’s Empire (1986); J. Raz, Ethics in the Public Domain (1994).

[27] Do not think this is so ludicrous. A prominent Canadian politician recently went on record suggesting that one way to limit pollution would be to stop breathing.

[28] Of course some would argue that academics that truly believe their work will be taken into account by the legislature, or by the courts, or even noticed by their colleagues are hopelessly naive. But such critics reject, in fact, the existence of prescriptive legal work. And what justification do these critics have then for the vast, vast majority of writing done in law with the stated goal of accomplishing just that – of bringing the law’s shortcomings to the attention of the legislature so that it might be amended, of bringing the decision’s flaws to the attention of the courts so that it might be modified and so on?

[29] This is the position taken by most contemporary writers on tort law, beginning with Posner and Calabresi. R. Posner, Economic Analysis of Law (3rd ed. 1986); G. Calabresi, The Costs of Accidents (1970).

[30] This is the position of Weinrib and Coleman. E. Weinrib, The Idea of Private Law (1995); J. Coleman, The Practice of Principle (2001).

[31] As I said the division of law into these levels of generality is solely for purposes of illustration and discussion.

[32] Imagine if you will the opposite. Of course, regimes have existed throughout history, and still exist, where the law was dictated and its discussion suppressed. Think of the social and political force that had to be exerted in order for change to be denied, however. It seems therefore that openness to change is a general feature of legal systems and the law, and that it is their default state in the workings of human society, so to speak.

[33] Dworkin, supra note 26.

[34] Id. at 188.

[35] J. Raz, Two Views on the Nature of the Theory of Law: A Partial Comparison 4 Legal Theory 249 (1998).

[36] There is a distinction between legal theory assuming that the law as it understands it is necessarily open to change, and between legal theory assuming that its understanding of law is necessarily open to change. Many theories assume the former while wholeheartedly rejecting the latter, although the latter is of great significance, as I shall discuss below.

[37] It is part and parcel, if you will, of its definition as prescriptive. One is of course free to disagree that legal writing can be divided into descriptive and prescriptive forms to begin with.

[38] Note that if society were in complete agreement over its purposes it would probably be possible to determine these effects with certainty. The law of such a uniform society would probably be limited to solving coordination problems, such as which side of the road one should drive on, and articles about the law would hold similar interest.

[39] With respect to rulings this has, of course, always been the case within the confines of the binding precedent doctrine of common law. Judges have historically elevated previous rulings or selected parts to fit their contemporary aims, and that is very much the case as far as legal writing in general is concerned as well.

[40] Hopefully this article will not prove an exception to the rule.

[41] This is of course the interesting question – what is the legal equivalent of quantum mechanics? Where does the Problem of Observation rear its ugly head? I attempt to answer this question below.

[42] Dworkin collaboration on the ‘philosopher’s brief’ is one example in which pedigree did not lead to the argument presented to the court having the desired merit.

[43] Again, that is why it is a Problem of Observation, and not a ‘Problem of Reality’.

[44] Not everyone would agree with all, or some, of these generalizations. Even those that disagree with them, however, have found them interesting enough to argue against. It is that these generalizations exist about law that is of importance, more than their accuracy.

[45] This is a somewhat simplistic characterization of Austin’s methodology.

[46] This is a somewhat simplistic summation of Hart’s argument against Austin. Hart, supra note 24 at 90-91.

[47] Raz, for instance, discusses the ‘essential features’ of law. Supra note 35 at 273 (footnote 38 to text). Dworkin discusses law’s ‘abstract’, ‘general’ and ‘fundamental’ points. Supra note 26 at 93.

[48] See Coleman, supra note 30 at 9. This methodology has existed long before it has been called ‘pragmatic’, however. Austin bestows the status of ‘law’ on norms simply because they are viewed as law by the public. Supranote 5 at 98. See also Hart, supra note 26 at 244. Waluchow, who is committed to a ‘descriptive-explanatory’ methodology although he does not discuss the role perceptions of participants play within it seems to distance himself slightly from Coleman’s ‘pragmatism’. W. Waluchow, In Pursuit of Pragmatic Legal Theory 15 CJLJ 125 (2002).

[49] Sometimes, of course, there truly is no law on a certain subject, simply because the subject itself does not exist. There was no internet law before the internet, no law on cloning or embryonic stem cells before they became parts of our life, and so on.

[50] Philosophy of law aims to change perceptions of law even when its stated purpose is to offer a general description of law, as is Hart’s stated purpose. Supra note 26 at 239-240.

[51] Hypothetically, of course, all the participants could agree, and share a perception of law. They could all become Dworkinians, or Razians, or Hartians or share some other perception. It is an interesting hypothesis to explore, especially in its implications for the prescriptive arguments these and other theories make, but it remains a hypothesis.

[52] Of course actual legal philosophy regularly combines both forms of argument and that in of itself is of no concern. The problem arises here since in my ideal triangle descriptive argument does not interact directly with legal reality, but serves as a basis for future prescriptive argument.

[53] Raz, supra note 26 at 199.

[54] Id. at 203.

[55] J. Coleman, Incorporationism, Conventionalism and the Practical Difference Thesis 4 Legal Theory 381 (1998); W. Waluchow, Authority and the Practical Difference Thesis 6 Legal Theory 45 (2000).

[56] Id.

[57] Raz, supra note 35.

[58] This goes back to the question what role perceptions of participants play within legal theory. Perhaps unsurprisingly Raz does not think they play a role in determining what the law is, while insisting that one perception – that of the law as an authority, is an essential feature of law.

[59] Such as, perhaps, Moore’s justification of positivism. See M. Moore, Hart’s Concluding Scientific Postscript 4 Legal Theory 301 (1998).

[60] One can of course continue to choose between theories on other grounds whether they are put forward from a descriptive or prescriptive perspective, and one of course always chooses between competing prescriptions and competing descriptions.


Return to Home Page