Knowledge, Opinion and the Audio Experience

Roy Harris

An audio experience is an effect of listening to recordings. Some examples include evaluation of the sound of stereo systems, changes in psychological and physiological states and enjoyment of music.

This brief essay is an attempt to show that knowledge is not a result of audio experiences. Rather, the outcomes from such experiences are either opinions, hypotheses or theories.

Consider the following definitions of knowledge and opinion:

Knowledge: Substantiated true belief

Opinion: Unsubstantiated belief

Both knowledge and opinion are based upon belief. Knowledge requires truth and justification (proof) and is held with certainty, as one cannot know that which is false. An opinion is held with confidence, based upon grounds insufficient to produce certainty. Hence it is a belief based upon the probability of truth. Knowledge, based upon analytic a priori propositions, axioms and definitions is attained using the principles of logic. Knowledge is necessary, being independent of the knower and experience. Opinions, hypotheses and theories, usually derived from observation, are contingent upon experience. Statements about stereo systems, e.g., are dependent upon a person having an audio experience.

The paths to knowledge and opinion are summarized below:

Reality: experience-->facts-->induction-->opinion, etc.-->probability

Abstraction: ideas-->propositions-->inference-->knowledge-->truth

Explanations of reality lack certainty, while propositions which are certain cannot explain reality, as reality is stochastic in nature. It would seem that knowledge and experience are incompatible.

The following tests may be used to distinguish knowledge from opinion:

1) Knowability

2) Validity

3) Acquisition

4) Proof

Knowability implies that information needed to acquire knowledge is both finite and available. That which is known should satisfy the criteria of content validity. The method(s) of acquiring knowledge must produce results which are consistent and reliable. Finally, the truth and certainty of what is allegedly known must be demonstrated.

I will now show that statements about recordings, components and stereo systems do not represent knowledge.

A simple argument will show that the "sound' of components and recordings cannot be known. Assume that there are CDs and components whose sound is initially unknown. After each CD has been played one notices differences in the 'sound' of the stereo system.

It can be proven mathematically that there is insufficient information to 'solve' for the 'sound' of each component and CD. Practically speaking, the quantity and quality of information makes it impossible to separate the sound of the whole, i.e., the stereo system, into its parts, i.e., the components and CDs. Therefore, the 'sound' of each is still unknown, even though all CDs have been auditioned. Hence, the test of knowability fails.

The sound of stereo system is affected by many variables, both intrinsic and extrinsic. Is the information necessary to describe the sound of a stereo system finite, or boundless?

If one auditions a stereo system forever, one cannot know if one has 'experienced' all that it and the recording have to offer. In addition, one cannot identify what has yet to be experienced or adequately describe what has already been experienced. Consider as well the fact that the 'sound' of stereo systems changes, as a consequence of oxidation, age, temperature, humidity, the AC and the psychological state of the listener. It is obvious that the 'sound' of a stereo system is unknowable.

The tests of acquisition, validity and proof will also fail if experience is the basis of knowledge.

Stereo systems can be experienced subjectively and objectively. In either case, the means for determining its 'sound' are unreliable. Ears, perception and memory are error-prone and inconsistent. The perceptions of experienced listeners may differ and the precision of test equipment may be questioned because of its degree of inaccuracy, which depends upon its 'standard error of the estimate'. Thus, the current methods of evaluating stereo systems do not produce knowledge.

Two types of validity are relevant to the sound of stereo systems, namely, content and description. Content refers to the terms used to evaluate the sound of stereo systems and description refers to the narrative used to describe each term. Stereo systems may be evaluated using terms such as timbre, tonality, dynamics and the artifacts air, separation and soundstage. Unfortunately, there does not exist a list of proscribed attributes so that the process for including some and excluding others may be arbitrary.

Setting aside the selection of appropriate attributes, there is another aspect of validity to consider, namely description criteria. Does a reference for a description of tonality, harmonics, etc. exist, especially when the 'sound' is created by recordings rather than directly by the instruments themselves? Since there are no established guidelines for what are and are not suitable attributes of the 'sound' of stereo systems and there are no formal criteria for specifying appropriate descriptions of the attributes, validity cannot be determined. Hence, the test of validity is inconclusive.

Proof of knowledge requires establishing the truth.

During the audition process, one may notice consistent sonic patterns. Making an assumption that the observed patterns will continue forever is inductive reasoning. Although one may believe that the 'true sound' of a stereo system has been revealed, propositions derived from facts are not logically certain, and as a consequence cannot produce knowledge. Empirically derived propositions such as the 'sound' of a stereo system is 'X' can never be proven true. One instance of disconfirmation is sufficient to disprove such a proposition. Whereas, an endless series of confirmations is necessary to establish truth which means that the truth can never be established.

I have tried to show that components and recordings are unknowable and that statements about stereo systems do not meet the requirements of knowledge. They are at best intelligent conjectures. If one acts on them there is a risk (probability) that mistakes and errors will occur. Unfortunately, there is no way to quantify the uncertainty and avoid disappointment.

Conclusions and Further Thoughts

Listening does not reveal the truth.

How does the absence of knowledge (truth) affect the following activities?

1) Purchasing components and setting up stereo systems

2) Diagnosing and remediating malfunctioning components

3) Analyzing the 'sound' of stereo systems

Regardless of whether the sound of a component is or is not known, its affect upon the sound of a stereo system is unpredictable. Therefore, ignorance of the sound of components is of little relevance to configuring stereo systems and may not impede the pursuit of one's sonic objectives. Troubleshooting (allegedly) malfunctioning components can be facilitated with the aid of logic, test equipment and an understanding of how components function. It is not necessary to 'know' that a component has been repaired or that a problem has been solved, as perception is reality, i.e., if it sounds OK it is OK.

The absence of knowledge does significantly affect the evaluation of stereo systems by introducing uncertainty and its consequences. The perceived sound of stereo systems is an unsubstantiated belief, i.e., an opinion. It is prudent to admit that one cannot know the true sound of a stereo system and act to minimize the effects of errors in judgment and the negative consequences of one's decisions. Auditioning a component in one's stereo system does not completely avoid the possibility of a surprise, as the sound of a component and a stereo system may change over time. One cannot be sure that a component is fully broken-in. Purchasing a demo after auditioning it in one's own system does not guarantee that its performance will stay the same forever. Although one may like what one hears at a point in time, sentiment may change at a later date.

There is a thriving market for used audio equipment. Perhaps used gear is the preferred mode of purchase because of the recognition that one can't be sure of the sound of components and stereo systems. Buying used equipment is a way to minimize a loss if an apparent synergy turns out later to be a mismatch. The recognition of change and the absence of knowledge may motivate behavior and partly explain the success of web sites such as Audiogon and E-Bay. It is obvious that the frequent sale of audio equipment represents a change in sentiment and or disappointment, both of which are unforeseen and a by product of the absence of knowledge. Fortunately, one does not 'change' one's spouse as often as one changes one's stereo systems. Divorce may be a result of lack of knowledge before and after marriage. If we learn from our human relationships we may not expect to ever know the sound of our stereo systems as we don't expect to completely 'know' our fellow human beings. Both are complicated and mysterious.

One can become a 'better' listener by improving one's focus and concentration skills and becoming aware of changes in physiological and psychological states which reduce the accuracy of perception. Increasing the accuracy of perception and memory through the application of appropriate forms of practice, in conjunction with improving the effectiveness of listening skills is likely to reduce the incidence of erroneous judgments. What ever measures are taken to make the audition process more effective, one is often faced with incomplete information and may consider the advice of 'experts'. Is such an approach a good idea?

'Authority figures' lose some of their 'authority' because their assertions are opinions rather than knowledge (truth). Professionals may be more talented, may have more relevant education and may be more experienced than non professionals. Does such superiority translate into 'better' opinions? The merits of opinions are difficult to assess. Unlike knowledge which can be tested, it is impossible to test an opinion before it has been followed. When decisions are made based upon the advice of professionals, the consequences in conjunction with the mechanism of conditioning and reinforcement determine reputations. Attitudes and sentiment are formed when advice is followed over a period of time. However, attitudes and sentiments may not be equally influenced by early and later experiences. It is possible that the first few experiences are more heavily weighted than those later on. The emphasis upon early outcomes may explain why someone may get 'hooked' on the advice of dealers and reviewers based upon a few early positive outcomes even though later ones are disappointing. Further opinions may be more difficult to change after some number of outcomes.

Confidence is the basis of action. Knowledge boosts confidence in one's ability to create favorable results. In absence of knowledge one can only rely on perception--one's own and that of others and objective measurement. When empirical propositions are followed by favorable outcomes, confidence is increased and the lack of knowledge may not be missed. If instead, perception is associated with negative results, one may regret the absence of knowledge.

Induction has and always will be the primary basis for decision making. If a component has a history of 'producing' a consistent sonic event after it has been inserted into many stereo systems, one may not hesitate to act on this empirical regularity without considering the possibility that in the future there may be an exception to the past. The formation of an expectation of future performance is based upon the mechanism of conditioning and reinforcement in the same manner as the formation of attitude and sentiment. The process is purely psychological and not intellectual.

Here are some ideas for operating in the world of probability:

1) Use the trend following approach when deciding when to follow suggestions of others. Maintain a record of positive and negative experiences. Try to devise a quantitative index, e.g., 'batting average'. Keep track of relevant facts as well as whether an outcome is favorable or unfavorable.

2) Ignore the advice/opinions of others. Trust your own ears and experience. Keep a record of results and quantify if appropriate. If unsatisfied with your own performance consider seeking advice.

3) Concentrate on the present. Examine your own ideas and perceptions. Consider the advice of others and analyze the merits of each position. Use deduction to ascertain the potentially best course of action independent of past results.