If you would like to download a pdf version of this essay, click here.

Abstract

   According to our metamethodology, physics itself has no aim.  Instead, we personally have to choose an aim in which we are interested.  We can then examine present and past investigations, with this aim in mind.  We find that there are some partially organised activities in which there is an explicit attempt to achieve the aim of {Truth} as a primary aim.  This aim is justified either directly, in terms of human values (interests), or indirectly, in terms of further aims - which themselves finally stem from values.

   Explanation and understanding are secondary aims, as is the search for generalisations. Predictive power is also secondary.  These aims are to dominated by Truth. The investigators intend to prefer true facts, with no plausible explanation, to a plausible explanation that we have reason to think is false.

   Any attempt to unpack "the aim of science is.." by analysis is incoherent, since it is people that have names, not abstractions.

Links

   To the Inductive Presuppositions that are automatically made by all investigators.

   To the Quarantining of Sceptical Doubt concerning these inductive presuppositions.

   To The Methods Of Science.  Once an aim has been chosen, we must identify the best available methods for achieving this aim.
 

   Are there any other characteristic aims?  We need to distinguish characteristic aims from (a) things which are desired, but which could be abandoned if they clashed with the aim of truth (b) things which are subsidiary, in that they are neutral with respect to the procedures of physicists.
   In category (a) we put: "Find realistic models"; this has come and gone several times in the history of physics.
   In category (b) we put: "Investigate areas which are potentially of interest or importance to people"; this is a sensible factor in investigation, but it does not affect the methods for the investigation of nature, only the direction of our original gaze.
   "Convenience" has been suggested by philosophers, and some philosophical physicists, but does not seem to us to be an aim actually used by physicists.
   For example, Fresnel argued that the analytical complication of his wave theory of light was irrelevant as a factor in assessing its truth; it was bad luck for us that the truth had turned out 'inconvenient' but, ultimately, so what?  He separated this from the logical simplicity of his theory, which he argued was a factor in assessing its truth. nature, he wrote, is not concerned about our analytical difficulties.  Convenience of use, he insisted, was not a reason for preferring the particle theory to the wave theory, when he was seeking the truth.
   As an aim, "Convenience" is inconsistent with the guiding theory of "Non-Anthropocentrism",  the theory that nature is so constructed as to take no particular notice of the existence of human beings.
   Few physicists, we judge, have followed the extreme view that, for example, a theory can be preferred, even if it is incoherent, clashes with other guiding theories, and has a selection of untidy special clauses for special cases, as long as it is easy to use.
   "Convenience" has, we think, mainly been supported as an aim by philosophers who have convinced themselves that physicists cannot sensibly attempt to have truth as an aim; for example, by those pragmatic philosophers who have claimed that truth is no more than what people find convenient to believe. This is far from the mainstream of physics and philosophy, so we are justified in not including it in our consensus.
   Physics could perhaps be done on this basis; it would be rather different from the activity that almost all physicists have engaged in for the last 500 years; if we thought that it had enough in common with 'physics', as a family resemblance, we could call it 'physics'; if not, we could call it something else.
 

   Understanding (explanation) does not have priority over truth; otherwise we would easily achieve it, by inventing any nice tidy scheme we fancied.
   It is loosely defined, on purpose, in order to be reasonable.  If physicists pinned down exactly what sort of understanding they wanted ("Or Else!"), they would be raising a specific guiding theory to the same level as the aim of truth; they would be saying, for example: "We want to find out the truth about the ways in which the personalities of the spirits which inhabit all the things in nature determine the apparent behaviour of things".  They would not be prepared to consider the possibility that nature was not animistic. This would only be reasonable, if a person's, or group's, desire for animism was stronger than their desire for understanding in general - in other words, if they said: "If the truth cannot be understood in terms of animism, we are not interested".
   We suggest that physicists take the same attitude, except that they do not specifically desire animism, or a final cause, or inanimate efficient causes; all they want is some kind of explanation.  Because of this broader desire, they are less restrictive on the kind of truths they are prepared to consider.  Nonetheless, they are still saying: "If the truth cannot be understood in some way, we are not interested".
   Is this more reasonable than the animists?  If we are purely seeking truth, then we would have to lay no restriction on it.  As soon as we lay a restriction on what we will accept, we are eliminating some possible truths.  The tighter the restriction, the more possible truths are eliminated.  But this is circular; if we want 'Truth + understanding in general', then making a rather specific restriction on the acceptable version of understanding is unreasonable. But if we do not want 'understanding in general', but instead want 'Truth + Animistic Understanding', then we are being reasonable in ignoring other possible ways of understanding nature, and single-mindedly pursuing our restricted vision.
   In other words, we suggest that physics is, as it happens, descriptively, characterised by a desire for 'Understanding in general'.  It is a reasonable activity for those that share this aim.  For those that do not, either because they do not care about the truth, or because they care about it only when it is more closely constrained, their own activities are equally reasonable.
   This conclusion, if correct, is important for the possibility of criticism - of comparison - of physics with other activities.  If, for example, a religious believer is living, and investigating, within a system in which their aim is "Truth + Christianity", such that the truth is permanently constrained within the limits of Christian beliefs, then they are acting reasonably within this system.  When we feel that they are not reasonable, it is because we are instinctively working in a different system, in which the aim is "Truth + some kind of understanding".
   But we have an advantage; our system includes theirs, even if they ignore our results. Christianity can be one of the specific kinds of understanding that we investigate.
   The vagueness of this aim is intentional; it is more negative than positive; it is the aim of finding the extent to which nature is not chaos; it is the aim of finding some kind of unity concealed in the diversity of phenomena. The particular rules of this 'not-chaos' are to be discovered. Each particular suggestion for a way of removing the chaos is a guiding theory: "Animistic Understanding"; "Teleological Understanding"; "Mathematical Understanding". The guiding theories come and go; the aim remains.
   An aim does not ultimately need justification; the strength of the desire provides its own justification, as we have argued above. If chaos fundamentally reigns in nature then we will call off the search for truth; we do not want to engage in this activity except on - to this extent - our own terms. This is our right, since it is our activity.
   Alternatively we could regard the aim as a desire to remove a discomfort - the discomfort produced by perceiving chaos in nature. This approach neatly avoids any question of specifying how this removal is to be achieved; the person on the beach can keep his options open by saying not "I want a Pepsi" but rather "I want something to remove my thirst". This is the aim; the Pepsi might be a possible guiding theory, but a drink we have never heard of might do the trick for us.
  

   It is an aim for the closely associated subjects of applied physics and engineering, but we are not convinced that it is part of the aim of pure physicists.
   We suggest, for example, that Balmer did not try to find a general law linking spectral lines so that he would possess the power to predict them in future.  Why would anyone care about predicting the wavelengths of light given off by hot gases?  Only, we suggest, engineers designing light-bulbs.  Balmer was looking for a general law because he thought that a general law would be a clue to the fundamental behaviour of nature (and because he liked numerical puzzles).  Since nature was fundamentally simple (Tg1), the law should be simple; generality is the most immediate form of simplicity.
   To try to find the way matter behaves, is implicitly to try to find the way that this matter, that matter, that other matter, indeed all specific bits of matter, behave.  Therefore the pure physicists' search automatically leads him to propose general laws - which are then very useful to the engineer, but were not devised particularly with this is mind.
   The same applies, in our view, to the aim of "manipulation of nature".  It is a side-product, of great human importance for good and evil, but it is not part of the characteristic aim.

   We therefore propose that the current characterising aims of the activity called 'physics' are, in order of priority: (1) Truth (2) 'Removal of Chaos'; in brief: 'Explanatory Truth'.  If a person is engaging in an activity which does not have Truth as its primary aim, then he is not presently classified as a 'physicist'.  If a person seeks the truth, but shows no sign of desire to 'Remove Chaos', if, for example, an experimentalist takes lots of photographs of bubble-chamber tracks, and presents them to a journal, or a theoretician presents a theory of elementary particles which is a long list of all known particles and their properties, they are not presently classified as a 'physicist'.  If a person has settled on a particular guiding theory, such as 'Realistic Models', which he wants and believes, this is OK, because the Tg are so far from the constraints of experience that the consensus on guiding theories is loose and liberal.  Indeed philosophical arguments - in the sense of natural philosophy - have always been relevant at the level of guiding theories (see Guiding Theories)

   No epistemologist has a monopoly on physics, its aims, and its methods.  Physics has no aim; it certainly does not have one which can be just laid down by someone, or derived from pure thought.  We can devise a model in which the aim is {1. Explanation; 2. True X}.  This model fits with some human activity, in particular with aspects of superstition, religion, and metaphysics.  We should compare it with a model in which the aim is {1. True X; 2. Explanation}.  This model fits with other human activities.  People who want explanations more than they want truth are investigators₁; people who want truth more than explanations are investigators₂.  Both can work rationally in their investigations.  Which one we personally value is another question.  And which one, or ones, fit closest with the investigations of the current consensus of physicists is another.
   I personally am interested in models which have {True X} as the priority, so I will continue to focus on these.

   Whether the aim includes explanation, and, if so, whether explanation requires realistic models, is an important issue in the debate between realists and anti-realists.  We therefore begin the discussion now; it is continued in the chapter on realism and antirealism.
   Realists may be tempted to argue that the aim of physics includes explanation, and explanation implies making realistic models of, say, the microworld, to explain the phenomenological gas laws and thermodynamics.  Hence - Hey, Presto! - antirealism is wrong.
   This neat argument is unsound.  Truth is the primary aim which we have chosen to use to characterise the activity called 'physics' - finding out the truth, the whole truth, and nothing but the truth, about nature - followed by the secondary search for explanation_x, loosely defined as the removal of arbitrariness.  We could, of course, decide that we are interested in another activity, physics_r, characterised by the search for truth, followed by the specific search for realistic models as explanations; but this would eliminate various investigators, such as Heisenberg, from the activity which we were hoping to characterise, such that we would realise that we were not intending to characterise it so restrictively.
   To put the point in another way, we could say that we valued this activity, physics_r, uniquely highly - and we would then be able to say that making realistic models is essential to physics_r.  But this is achieving nothing, because we realise that we have merely used a persuasive definition - we could have defined 'physics_x' any way we wanted, and so can our antirealist critics.  We are actually trying to capture some pre-existing classification, some concept, not merely invent one of our own.  
   To capture this concept, we find that we need to return to the broadly inclusive aim, where truth is primary and explanation_x is secondary.  But Van Fraassen is correct that this implies that the antirealist interpretation of 'explanation_x', in which it involves the reduction of observable phenomena to simple mathematical structures, is equally as acceptable as the interpretation which involves using realistic models of  low-observability entities.

   In brief, realists and antirealists cannot use the aim of physics as a simple stick with which to beat their opponent.
  But, as we argue in the chapter on realism and antirealism, the overall aim of truth provides a weighty argument for attempting realism (unless the antirealist can prove that it is impossible).  The whole truth will include the truth about low-observability aspects of nature.  Therefore an antirealistic physicist who refuses to seek the truth about the microworld is not doing his duty.
   An antirealist could respond by saying that he is at liberty to restrict the aim of physics to seeking the truth only about aspects of nature which are highly observable.  Once again, this is true, but it is a persuasive definition.  The easy victory is Pyrrhic, because the antirealist has failed to capture the pre-existing classification of 'physicist'.

   Suppose that the antirealist responds that he is not just arbitrarily inventing convenient definitions of 'physics_x' - we cannot obtain the truth about low-observability aspects of nature; evidence, acting indirectly,  is not capable of achieving it.  Therefore the realist aim is not rational.  This response is important, but it is no longer an argument based on juggling with the aims of physics.  Realists and antirealists need to accept that the activity in which they are both interested - which they both value - is characterised by the primary search for truth, and the secondary search for explanation (loosely defined).  The resolution of their disagreement needs to be achieved within this constraint.

Bacon (The New Organon Bk.1 LXXXI): "The true and lawful goal of the sciences is none other than this: that human life be endowed with new discoveries and powers".  This in turn is to be achieved by "the severe and rigid search after truth".   In Bk. 2 (p.267 LII) he writes of the aim as "an improvement in man's estate, and an enlargement of his power over nature".

Maxwell (1984) has argued that little sense can be made of physics unless the aim of physicists is regarded as Explanatory Truth rather than just Truth.
 

Poincaré (1905) (p.131) "Every law is held to be simple till the contrary is proved". We suggest that this is partly because we want it to be true and simple; it is also, as Poincaré says repeatedly, because we have evidence E that Tg "Mathematical Simplicity" is true.
 

Newton-Smith (1981) suggests that the aim of physicists is Truth + Explanation.

Dien Ho, at the BSPS conference in July 2001, presented some arguments heading towards this conclusion.  He was arguing that the logicist research programme of analysing explanation, broadly proceeding along lines established by Carl Hempel, had been degenerating for some time.  What we call an 'explanation', he argued, is always something which contains an epistemic (his word) component; the mere existence of a logical relationship between sentences is not sufficient.  Suppose, for example, that we agreed that physicists devise covering laws, universal generalisations from which particular observation statements can be logically deduced.  The question immediately arises "Why?" - why are they interested in this logical relationship rather than others?  The natural answer is that this is one of the ways by which they are trying to satisfy their desire for knowledge.  But this leads us towards the suggestion that it is this desire that is primary, and that satisfying it by devising covering laws is secondary.  It also implies that there may be other ways of satisfying it.

I read Burtt (1967) as an essay on this theme.  Present physics should takes its place as one in a sequence of human attempts to understand nature - although it is very different from, say, investigations based on animistic explanations, fundamentally both are part of the same human project; both have something fundamental in common, which is that both are attempts to seek explanatory truth - to reduce the distressing apparent arbitrariness of our experience.