Empiricism, Explanation and Rationality

An Introduction to the Philosophy of the Social Sciences

Len & Roger Doyal & Harris

• Publisher: Routledge
• Year: 2013
• ISBN:
  9781135028695
• Edition: 1
• Topic: Philosophie
• Subtopic: Histoire et théorie de la philosophie

Originally published in 1986. All students of social science must confront a number of important philosophical issues. This introduction to the philosophy of the social sciences provides coherent answers to questions about empiricism, explanation and rationality. It evaluates contemporary writings on the subject which can be as difficult as they are important to understand. Each chapter has an annotated bibliography to enable students to pursue the issues raised and to assess for themselves the arguments of the authors.

1
THE PROBLEMS OF EMPIRICISM

Traditional ideas about scientific method have been the target of much recent criticism. This has been directed particularly against empiricism — a doctrine which proclaims that all knowledge ultimately originates in experience. Empiricists argue that the final arbiter of any dispute in science must be the evidence of observation. Where observation is unable to play this role then dogma, speculation or superstition — anything but science — are said to be the result. All those who do not abide by the precepts of empiricism are thus threatened with excommunication from the bosom of science. This view of science has come under increasing attack in recent decades. The critics of empiricism claim that it never was and never could be the method of science and that conscious or even unconscious adherence to its principles can retard progressive scientific discovery. Indeed, there are some who argue that there is no such thing as scientific progress and that the importance of observation in scientific inquiry is minimal. Empiricism is therefore at the centre of fundamental disputes over the methods and, more important, the authority of natural and social science. This chapter will provide an introduction to these disputes. The first two sections argue that observation cannot be viewed as the foundation of scientific knowledge. However, the third section makes it clear that empirical research is still vital for scientific inquiry but that its importance cannot be understood by an empiricist theory of knowledge.

Crude empiricism and its problems

Social scientists, like everyone else, are philosophers of knowledge. In other words, they all have theories of knowledge — ‘epistemologies’ — of which they may not be clearly aware, but which nevertheless structure their attitudes about what is and is not ‘scientific’. Such theories offer solutions to three key problems. What is the origin of scientific knowledge? How is the merit of such knowledge established once it has supposedly been created? What is meant by scientific progress and how can it be identified? In future, we shall refer to these problems as those of scientific creativity, scientific assessment and scientific progress respectively. Epistemological theories are important because they are used to decide which sorts of explanations of the natural and social world to take seriously. They also influence criteria of good scientific practice and training.

Empiricism in its crudest form is probably the epistemology which is most generally accepted by people without philosophical training. It embodies the most common beliefs about successful science and scientists and is implicit in the images employed in the media to depict them. Crude empiricism assumes that the scientist is a sort of spectator of the object of inquiry. In other words, reality is presumed to exist externally to scientists and its structure and content are seen to be independent of their beliefs or desires. The problem scientists face is thought to be analogous to the popular conception of the problem facing the good journalist — how to make contact with reality which is not distorted by preconceived ideas about what it might or might not be like. The scientist is thus viewed as a subject who is attempting to understand an object and is trying to be objective by eliminating the bias that could lead to inaccuracy. The result of such understanding is said to be scientific knowledge which can be accepted as certain or at least very highly probable. If contact with reality is distorted by subjective preference the resulting understanding is relegated to the status of mere belief or opinion. It may be interesting, colourful or moving but it will not be science.

It is all fine and good to present the prospect of such undistorted contact but how is it to be achieved? For crude empiricists, scientific inquiry consists of the undistorted recording of particular observations. Such observations may be gained, for example, from experiments or surveys — both methods of collecting empirical data which have been devised to facilitate the process of discovery. This picture of the successful scientist with a passively receptive mind has been likened to a ‘blank slate’ on which nature writes its experiential message. Provided the data of such experiences are not literally misperceived, again through bias of some kind, the message can be securely relied upon. Reality will have been explained.

Thus if you want to be a good scientist it follows from crude empiricism that you must begin by ridding yourself of all preconceptions about what you are going to study. Indeed, this is one of the main things a good scientific education is thought to achieve. You will then be in a position to collect particular observations which you hope will reflect the regularities of nature or society relevant to the problem you wish to solve. When you have collected what you consider to be enough evidence, you may develop an explanation of the problem by generalizing from it. If your explanation is firmly grounded in such evidence, it can be assumed to be certain. The less firmly grounded, the less probable the explanation. This form of empirical reasoning is known as induction. It is the method proposed by crude empiricism to distinguish scientific inquiry from non-scientific speculation and until the 1960s was regarded in this way by most scientists and many professional philosophers.

For example, suppose you are interested in the causes of unemployment in inner cities. Crude empiricism dictates that you must begin your investigations without anticipating their results. In a variety of different ways you will systematically collect as much evidence about such unemployment as possible. You might start by observing those things which consistently seem to be correlated with it (e.g. poor investment, poor education, bad housing, certain types of psychological attitudes among the unemployed). But to arrive at a correlation which you believe to be causal rather than accidental, you will need to devise experimental ways of eliminating those factors which sometimes just happen to accompany unemployment but cannot be said to cause it. Much literature about research methods in the social sciences contains detailed descriptions of experimental designs which attempt to achieve this aim. After carrying out such experiments you might conclude that some factors (e.g. bad housing) do not cause unemployment since you have observed that they do not always accompany it. Or you might conclude that other factors (e.g. poor investment) are causal because on the evidence they and the employment trends are correlated.

Once a factor is identified which on the available evidence the researcher believes to be causally connected with the events to be explained, crude empiricists argue that the next stage of the inquiry involves a process of generalization. By inductive reasoning from particular evidence to general conclusions, an explanation is formulated which purports to apply to all instances of these events regardless of when or where they occur. This will not only facilitate understanding but also enable predictions to be made. Suppose, for example, you conclude that lack of investment is the primary cause of all unemployment in inner cities. It will thus be possible to deduce from this generalization that wherever there is a significant lack of urban investment, it will be followed in proportion by a rise in urban unemployment. Then you can check to see whether or not this is the case. If not, the crude empiricist would conclude that the inductive method of generating the hypothesis in question had not been properly followed. Much the same process goes on in everyday relationships. When you think that your beliefs about someone are justified by your past experiences of them, you predict their responses to new situations by deducing what follows from these beliefs. In short, for crude empiricism, induction is the method of consolidating the observational link between scientists viewed as inquiring subjects, and reality viewed as the separate object of their inquiry. By contrast, deduction embodies the method by which knowledge that has been inductively generated is applied and extended to other empirical situations not yet observed.

If scientific work is carried out in this mechanical fashion, then solutions to the epistemological problems of scientific assessment and progress are said to follow automatically. Proper use of the method of induction deals with assessment for it is taken to lead inevitably to scientific explanations which are either certain or probable to a degree which can be calculated with some certainty. For example, a crude empiricist might argue that the physical law which states that gases expand when heated is certain because it has been proved by the particular observational evidence from which it was induced. Or it might be argued that though absolute certainty is unobtainable, it is possible to calculate the probability of the law on the basis of all of the known evidence which confirms it. In any case, it is assumed that if the inductive method of theory construction is correctly applied, the problem of scientific assessment will take care of itself. And if this same method is followed by successive generations of scientists, then progress will also be inevitable. According to crude empiricism, the history of science is the ever-increasing accumulation of particular observations and of general theories systematically generalized from them. It also involves the accumulation of successful experimental methods developed to extend observational capacity and to ensure observational accuracy. Scientific progress is therefore portrayed as cumulative. The accretion of facts by induction takes place step by step, gradually yielding increasingly reliable knowledge.

The theory of crude empiricism is intuitively appealing because on first hearing it sounds like a sensible way for inquiry to proceed. Moreover, its conception of human beings and their place in nature seems to be supported by natural science. Humanity is believed to be part of nature and as subject to the same laws that govern every other part of the material world. Thus all the information that you or any other organism receives from the surrounding world must derive from some physical process to which you are subjected by your environment. The senses are located in and depend upon those organs of the body which are especially sensitive to particular effects of the environment impinging upon them (e.g. light, heat, sound, pressure, etc.). Two factors supposedly determine the capacity of an organism to learn from such stimuli. First, it is assumed to be sensitive enough to discriminate between small changes in its surroundings over a wide range of possible stimuli. Second, the organism is believed automatically to correct the flux of incoming data in a manner which corresponds with the regularities of nature. This is why, for example, someone who has suffered brain damage may lose the capacity to experience the world in the ‘normal’ way required for scientific investigation to proceed. Learning on this model must be inductive, it is argued, since all the organism learns from are particular sensory inputs. Therefore, the growth of science is assumed to be a natural causal process akin to the rudimentary learning processes that occur in simple animals. While humans may differ from them in degree, the basic inductive picture of the accumulation of knowledge remains the same for both.

How then does the crude empiricist explain what are now regarded as the mistakes of the scientific past? Previous errors are, again, basically to be understood as the result of the improper application of inductive method. The scientists involved simply distorted the reliable observational link between knower and known because of their prejudice, bias, superstition or whatever. It follows from crude empiricism that the book of nature or society is written in observational terms and can be correctly read by all who open their minds, reasoning carefully by induction from the particular to the general. Of course it would be granted that primitive experimental technology can also limit scientific achievement. However, bias or avoidable error is held responsible if scientists are blind to the experimental potential of existing technology or to the need for new and better instrumentation. For example, crude empiricists like to recount the oversimplistic tale of Galileo's opponents who refused to look through his telescope because their experiences might have conflicted with their prejudices. Similar tales and images dominate the presentation of science in school and in the media and their influence explains why crude empiricism is such a widespread doctrine. Yet it is now rejected by most natural and social scientists. Among philosophers of science it has long been regarded as foolish. Why?

There are four major problems with crude empiricism: selectivity, certainty, error and interpretation. We will begin with the problem of selectivity. You have seen the background to the argument that preconceived ideas — beliefs which are held prior to empirical inquiry — are anathema to any scientific process of discovery. They supposedly lead to error because it is argued that the truth will reveal itself only through perceptions undistorted by preconceptions. This is why crude empiricists claim that the scientific method entails reasoning from particular observations to general theoretical conclusions, and not vice versa. But it should be clear that any scientific investigation must begin by picking out observations which are relevant to the problem at hand from the infinite number of others which are irrelevant. Scientists do not collect evidence in a conceptual vacuum and they cannot observe everything. When a specific inquiry begins, much potential evidence has already been deemed inappropriate. Thus rigorous selection has already taken place, even though the inquirer might not be consciously aware of it.

To continue with our previous example, if you were attempting to explain urban unemployment it is unlikely that you would regard the number of bricks, manhole covers or pigeons in unemployed areas as relevant evidence. It is equally unlikely that the colour of hair, the number of pets or the type of toothpaste of the unemployed would strike you as important. Selection is inevitable and since it is particular observational evidence which must be chosen, scientists cannot simply be reasoning from the particular to the general when they select. Therefore, their general theories about what might be the case must lead them to look out for particular observations which they expect to be the case. Such theories constitute expectations both about what will happen in specific situations and about what should be defined as ‘natural’ and not worth attention. This is why scientists are surprised when they are mistaken. Yet since they are prior to observations deemed scientifically relevant, it is precisely these expectations that the crude empiricists must incorrectly reject as impeding the progress of science. So, contrary to their doctrine, order and understanding are not imposed on the passive inquirer by the data of observation. Rather, they are actively created by the inquirer selecting from the otherwise indiscriminate input of the senses.

The problem of certainty is intimately related to that of selectivity. It arises because when scientists develop explanations of events, their aim is to show why those events (rather than any others) occur when and where they do. But in all fields of science it is necessary to go beyond a simple recounting of the event to be explained. Any understanding of the expansion of a given volume of a gas when it is heated, the generation of electricity from a battery, the cognitive development of a child or the defeat of the Spanish Armada requires more than a description of what is already known! The more empirical situations there are to which explanations apply, i.e. the more universal their application, the more worth critical exploration they will seem. For example it is clearly more challenging theoretically to argue that all gases expand whenever and wherever they are heated than to suggest that such expansion only occurs in the labs of one university on Friday afternoons. If Karl Marx's theories of capitalism were restricted in their application to London during the period when he worked in the British Museum, they would have attracted little interest. And so on. Yet this need for explanatory generality poses a problem concerning assessment for crude empiricism. For the more the scientific attractiveness of an explanatory theory depends on its general applicability, the more difficult it will be to demonstrate the certainty or high probability on which the doctrine lays so much stress. Indeed, if any theory really does purport to be universal in its application, its proof will be impossible. There will always be the possibility of discussing conflicting empirical evidence which will not have been realized either for reasons of practicality (e.g. acquiring the funds to do all the necessary research) or of principle (e.g. there are parts of the universe which we shall never reach to do any research).

In the next chapter we shall look further at what more might be involved in explaining something, over and above merely recounting what happened. Suffice it to say at this point that the need always to go beyond a mere report of events in order to explain them means that the truth of explanations is always under determined by the evidence available. Everything a theory explains could be evidence for its truth, but in order to account for this evidence, the theory must go beyond it. Again, this is why a theory can never be fully supported or proved by the evidence of what it can explain. It is in this sense, for example, that even assuming that all of the available evidence supports Marx's theory of capitalism, it cannot be said to be certain — any more than the evidence that you have never been run over by a car proves that you never will be. Or to use a famous, though hackneyed, philosophical example, no matter how many white swans you may have observed, this does not prove the contention that ‘All swans are white’. Ask any Australian! In short, although many interesting explanatory theories are taken to be universal in intent, it does not follow that they will be successfully universal in application. Because of this problem of underdetermination, theories are — like their originators — fallible. Accepting such fallibility does not mean denying that some theoretical beliefs are better than others. It simply means that it is necessary to find some other sense of ‘better’ than ‘certain’.

The third problem with crude empiricism — error — concerns this essential fallibility of scientific knowledge. You have seen how it is argued that science proceeds through inductive reason from particular, correct empirical data to general, correct theoretical explanations. That is to say, crude empiricism claims that scientists learn through not making mistakes from the very beginning of their inquiries. But can this possibly be so? You already know that most of your own important learning experiences have involved getting it wrong, making mistakes. Learning occurs against the background of a variety of expectations which are accepted as correct but are not necessarily recognized for what they are. These expectations are analogous to theoretical beliefs in that you use them to predict what will happen in specific circumstances in the natural and social world. Recall what has already been said about selectivity: when you have meaningful learning experiences, it is because you discover that some of these expectations are false. This occurs when what you predict to be the case turns out not to be so. Such mistakes constitute the origin of most genuine motivation to learn and to attempt to make new discoveries. You do not feel that you need to bother when your experiences continue to confirm what you already believe to be correct. Since the process of discovering empirical error involves reasoning from general expectations on the level of belief to particular expectations on the level of observation, this must mean that learning does not occur inductively as crude empiricism claims. To assert that one learns through making mistakes may sound like a cliché. However, in relation to the psychological and educational impact of crude empiricist conceptions of learning, it is revolutionary. Think, for example, in light of the extent to which a fear of making mistakes can often cripple individual participation in education, of how liberating it would be to recognize the importance and inevitability of error in the learning process.

The fourth problem facing crude empiricism is that of interpretation. We have shown how ...