This is your philosophy

Everybody – including scientists – makes seat-of-the-pants practical judgements about what’s real and what’s not. The common-sense assumptions underlying these judgements can be unrecognized, inconsistent and even untenable; they can be home-grown, inherited and absorbed from others. But when someone is engaged in an activity as complex as science, it is almost impossible to avoid making such practical judgements. No matter how implicit and readily revised these judgements may be, they are based on preconceptions of what the world consists of and what the world’s most important distinctions and categories are – in other words of how it all hangs together.

Professional philosophers analyse these preconceptions and up the ante on them. They formally rework the assumptions into consistent, fully articulated and intellectually supportable positions. They then give them names, such as realist, antirealist, critical realist, constructivist, hermeneutical realist, and so on. To qualify as a philosophical position, it has to be advanced in clear words, articulated in appropriate detail and depth, and be defensible against criticism when scrutinized in a philosophical peer review.

Why philosophy shouldn’t be avoided

I’ve often heard scientists call philosophical attention to their field irrelevant at best, and confusing and destructive at worst. Indeed, many scientists advise that philosophy should be avoided altogether. Steven Weinberg, for example, named a chapter in his book Dreams of a Final Theory “Against the philosophers”. Murray Gell-Mann, meanwhile, has remarked that philosophy “muddies the waters and obscures [the theoretical physicist’s] principal task, which is to find a coherent structure that works”. He then added that having a philosophical bias may cause a physicist “to reject a good idea”.

But such reactions misconstrue philosophy, however much they may have been triggered by the excesses of philosophers themselves. Scientists cannot avoid making judgements about what is real and what is not, and philosophical analysis seeks to expose and clarify this process.

I’ve also heard that science inclines its practitioners towards a specific philosophical position. Scientists, it is said, tend towards realism because it makes them better scientists – a conviction that has also influenced philosophers. When Ian Hacking, for example, once asked a physics colleague what he was doing, the physicist replied that he was “spraying photons”. Impressed, Hacking wrote: “From that day forth I’ve been a scientific realist. As far as I’m concerned, if you can spray them, then they are real.”

In his book Faith, Science and Understanding, physicist-turned-Anglican-priest John Polkinghorne remarked that “virtually all scientists” – including himself – adhere to a brand of realism known as critical realism. A reviewer in Physics World, who doubted Polkinghorne’s bold assertion, later suggested that I poll readers, hoping to elicit information to settle the issue. I therefore carried out a survey in which I listed a number of different items and asked readers to say whether or not they considered them to be real things, or whether they were unsure (Physics World October 2001 p18). Having received more than 500 replies, the statistics (see tables) do indeed cast doubt on Polkinghorne’s claim.

Different brands of realism

Realism is the view that things in the world exist that are not of our own making – independent of human perception and thought – and that scientific theories are true if they faithfully correspond in some way to these things. Indeed, one of the strongest lines of reasoning for realism, termed the miracle argument, is that – as the philosopher Hilary Putnam once put it – “realism is the only philosophy that doesn’t make the success of science a miracle”.

But a tension lurks in the meaning of “real”. People who insist that real things are the sort incorporated into the fabric of human experience – i.e. order as human beings experience it – adhere to a brand of philosophy that has been called phenomenal realism. They view what things are like, apart from that experience, as unknowable.

Meanwhile, those who argue that we can, in fact, come to know something about structures that underlie the fabric of human experience – i.e. nature’s order – adhere to what is known as noumenal realism. These structures may be material (Democritean) or abstract and formal (Platonic), such as (in this view) electrons and protons. What is truly real – for noumenal realists – are particles and forces, say, rather than sticks and stones. In this view, what human beings experience is not of reality as such, but rather the means and clues by which we can know underlying, fundamental structures that are not given in experience.

Different branches of realism evolved partly in response to this tension. Critical realism – Polkinghorne’s pet idea – is a term that covers several different perspectives, combining aspects of phenomenal and noumenal realism. What these perspectives share in general is the view that what we directly perceive and know is not the real object itself – an electron, say – but rather a sign or datum by which we can infer the existence and properties of the object. These inferences, however, may often fail to capture the object’s details and even its essence.

Polkinghorne’s views fit this loose orientation. By realism, he says he means that science “faithfully represents” the world, giving an “increasingly verisimilitudinous account of what the physical world is like”. This does not mean that Polkinghorne thinks that the world is intuitively graspable – after all, he admits that it contains strange entities like quarks and superstrings, and that mathematics is its natural language. And to explain what he means by the word critical, Polkinghorne says that “scientific understanding is not just read out of nature but…attained by a creative interpretive process”.

Different brands of antirealism

But a strict definition of realism quickly runs into trouble from several directions. One is that scientific theories are often about prediction and control, not explanation and portraiture. Another problem is that scientific theories and practices change in non-trivial ways, forcing adherents into the awkward position of having to dismiss most past science as false, and of having to prepare to say the same about present-day science should future theories of a radically different character replace those of today.

Yet another difficulty with realism stems from the fact that much of what scientists study is a product of their theories. As J Robert Oppenheimer once put it, 20th-century science “forced us to reconsider the relations between science and common sense, [and] forced on us the recognition of the fact that we were in the habit of talking a certain language, and using certain concepts did not necessarily imply that there was anything in the real world to correspond to these”.

Such counter arguments led to the development of so-called antirealist positions. Empiricists are antirealists who argue that the purpose of scientific theories is not to correspond with some basic structure of the world, but to organize data. Constructivists stress the extent to which concepts and theories are socially constructed – much like the rules of a game – rather than discovered in nature. Operationalists say that reality can be attributed only to observable elements by virtue of operations that can be performed on them. Instrumentalists hold that concepts and theories are tools to be evaluated for their usefulness rather than truth value.

The most significant challenge to realist positions is to account for non-trivial theory change. The biggest challenge for antirealist positions is to explain the stunning success of science, why some theories are so hard to change or abandon, and why we have such little control over which theories seem right.

Hermeneutical realisms

Yet another alternative philosophy of science – besides realism and antirealism – is hermeneutical realism, which comes from the Greek hermeneutikos, meaning interpretation. Hermeneutical realists think that existing forms of realism and antirealism start off on the wrong foot by presenting us with a dualistic series of forced options. Either theories (inside us) represent things in the world (outside in nature) or they do not. Either we (inside) have direct access to something independent of us (outside), or we do not. Either theories are successful because they contact the real world (over there) or because they order the data (in here).

Hermeneutical realists do not think that we are faced with those choices. They see the real as the relationship between ourselves and our surroundings, not as a subset of the things that we find in those surroundings. In other words, hermeneutical thinkers do not begin with isolated knowers and knowns. Instead, they say that it is impossible, in that relationship, for us to peel away theorizing and what there is from each other. They then examine the practical judgements involved in that relationship.

Hermeneutical thinkers insist that to perceive something as real – rather than illusion or error – does not merely involve data taking (phenomenal-realist style) or deciphering a clue that tells us something about another, somehow more real world beyond what we experience (noumenal-realist style). Perception is neither an event in the head nor code-cracking. Instead, we perceive something to be real – in science as in everyday life – if we find it to behave in a predictable, law-like way within a particular background context or horizon, fulfilling (or not) our expectations.

When we see a profile of George W Bush on the street, say, we perceive it to be a cardboard profile or the real person by walking round to view it from other angles. For the hermeneutical realist, scientific data (about electrons, say) are like profiles of the real George W Bush. Data provided by different experiments – or by variations of the same experiment – designed by the same theory and within a common (scientifically controlled) background context are found to cohere in a predictable way as the object is sampled from different directions, allowing us to “read” these samplings as profiles of the same object.

As we develop and improve empirical practices (the instrumentation and techniques for handling electrons) and the background horizon (electromagnetic theory), data and object will appear differently. Four decades ago, the reality of the big bang was considered to be confirmed by the measurement of uniform background radiation; a decade ago, it was considered to be clinched by the measurement of deviations from uniform background radiation.

Hermeneutical realism is not a constructivism because the objects that we perceive to be real exhibit invariances that are not under our control. Yet it accounts for non-trivial changes in the history of science. In the hermeneutical perspective, scientific knowledge evolves from – and at the same time transforms – historically inherited practices and judgements. There is something unavoidable about this insight; if it didn’t happen, science would be impossible or trivial.

Your views

I’d like to thank all of the 534 people who replied to the poll, some of whom wrote lengthy accompanying letters. Some Italian physics students at the University of Bari even gave it to their professors to fill in. I’m gratified that many participants found the poll harder to answer than they thought it would be – including the reader who had originally allotted himself, speed-chess style, five seconds for each answer.

Some people said that they changed their minds, or grew more uncertain, or realized that their answers were inconsistent, while doing the poll. (It made them think!) A handful of people even stopped in despair partway through. Yes, the phlogiston button was broken and viscosity was misspelt. And to those who complained that “it depends”: precisely! The philosopher’s task is to discover the variables and how they affect the outcome.

The results reveal a full spectrum of positions. A handful of people were fellow travellers of Weinberg and Gell-Mann, including the person who called the poll “a trap for the unwary” and the physicist who archly informed me that “in the last 40 years, [you] philosophers still have not found a way to ask a physicist a ‘real’ question”. Each, with admirable consistency, returned the poll conspicuously blank. The person who wrote that “a scientist who is not a realist is either lying or incompetent” was a realist fundamentalist.

A few people said they were operationalists or constructivists, but most who explicitly labelled themselves said they were realists. What I learned from the poll was that it was difficult to tell from people’s answers what kind of position they adhered to. Sometimes it was easy. I’d classify as a naive realist the person who simply wrote “I’m a realist” on his poll, and ticked off everything accepted by contemporary science as real and everything else as not real.

I’d identify as phenomenal realists those who shared the views of the person who wrote: “I take objects that I can ‘touch’ directly or not, as real. I take concepts as unreal.” Such people were usually (though, surprisingly, not always) among the 10% or so of respondents who rejected things like quarks and electrons.

People whose working definition of real things was anything that – as one respondent put it – “had an existence in the real world even if not observed” were realists in either the phenomenal or noumenal sense. These people were among the 43% and 32% who believed that hallucinations and after-images are not real.

Those who said that real things were made of matter or had mass were Democritean noumenal realists – and were therefore among the third or so of respondents who rejected the reality of the gravitational constant G, as well as numbers. The physicist from the Lawrence Livermore National Laboratory who wrote: “I think that what I’d call ‘real’ are concepts that work,” was an instrumentalist. So were the large fraction of people whose criterion for reality was usefulness, such as the person who measures the reality of an idea by “how well it can predict the outcome of a bet”. A large number of people said that the real was the measurable, which could be a sign of realism, operationalism, or hermeneutical realism.

The large overlap between those who believed in the reality of colours (50%) and light waves (68%) is revealing. When it came to saying what was real about light, these people evidently did not think that they had to choose between something necessarily experienced and something abstract and third person. They were comfortable with the fact the phenomenon of light appears differently in different background contexts.

Another person who felt no need to distinguish between what is experienced and what is abstract was the retired scientist who told me that kinetic energy is real “when considering the effects of bullets and sledge hammers, but…not real when considering changes of frames of reference”. And so did the people who worried whether quarks, wavefunctions and electrons will go the way of phlogiston.

Names like quark, electron and even atom are ambiguous, another reason why it’s difficult to tell a person’s philosophical allegiance from their choices alone. For in a distinction insisted on by hermeneutical realists, these names can be taken to refer to abstract terms in a theory – like notes on a musical score – or to actual objects manipulated in the laboratory, like a note in a concert hall. It’s interesting, therefore, that more people believed in the reality of atoms and electrons (84%) than quarks (68%), even though all three are accepted theoretical elements. Evidently the fact that quarks cannot appear singly makes them unreal for some people.

Terms like the Bohr atom, the Copernican system, and so on, are also ambiguous. Those who said that these were real (or not) because they were useful (or not) were instrumentalists, while those who said they were real (or not) to the extent that they correlated with measurements were operationalists. And those who said they were real (or not) because they represented the world with reasonable accuracy were realists.

My overall impression from the poll results and associated comments is that many (and perhaps even most) respondents could well accept the basic tenets of a Polkinghorne-style critical realism – namely that the purpose of a theory is to represent, that the world includes abstract and sometimes counterintuitive objects, and that mathematics is its natural descriptive language. However, there are sizable fractions of the sample who clearly disagree with such a view. Some people, for example, insist on tangibility as a criterion of the real, while others think that models and theories are real only if they are useful or operationally successful, rather than descriptive.

Still more significantly, a large fraction of respondents cannot be classified as critical realists because they recognized, while answering the poll, that their answers were philosophically indeterminate. Indeed, the most heart-warming letter I received said: “At the end of [your original] article, you said that a low response will indicate either that you have no readership or that scientists don’t care about the issues raised. After 48 hours of discussions we have to suggest a third category – those who would like to reply but in attempting to answer the questionnaire have found their ‘gut’ philosophical position to be wholly inadequate and inconsistent.” Her poll, too, was blank – but it seemed a product of a sensitivity to the seriousness and significance of philosophical issues rather than a repudiation of them.

The critical point

Einstein once wrote that, to the philosopher, the scientist appears as an “unscrupulous opportunist”, for “he appears as realist insofar as he seeks to describe a world independent of the acts of perception; as idealist insofar as he looks upon the concepts and theories as the free inventions of the human spirit (not logically derivable from what is empirically given); as positivist insofar as he considers his concepts and theories justified only to the extent to which they furnish a logical representation of relations among sensory experiences”.

Einstein implied that one need not have a consistent philosophical position to be a good scientist. Philosophers may suppose that scientists, being rigorous and conceptually savvy, must have fully worked out positions. Scientists, meanwhile, may assume that the views they hold about reality correspond to the positions that philosophers have found to be the most rational. But fully articulated positions are only for those being needlessly rigorous and consistent.

Furthermore, it would be difficult to translate everyday assumptions into such positions, just as it would be difficult to translate an average person’s political and religious views directly into a systematic political programme or theology. Even philosophers may not commit themselves to worked-out positions, for philosophers are less holders of positions than examiners of them. Nevertheless, as in politics and theology, a position does not need unconditional subscribers for there to be value in formulating and examining it.

What is this value? A fully articulated position would relate the kinds of practical judgements and decisions (and their underlying assumptions) that scientists make to those that human beings make in other kinds of activities. A convincing account of this relationship would make it possible to critically examine the stories – told by scientists and science’s critics alike – about what science is and what it does. These stories are not harmless. They can provide false impressions of what science can and cannot accomplish, and can promote a false sense of the promise and danger of scientific activity in human affairs.

By articulating the relationship between scientific practice to other kinds of human activities, a fully articulated philosophical position would make scientific judgements and decisions appear less abstract, strange and arbitrary to outsiders. This would help to re-establish a dialogue between the scientific community and its clients, supporters, academic interpreters and the public at large. There is a danger – not only to science but also to the public – if this dialogue breaks down. The so-called science wars are only the most recent manifestation of the breakdown of this dialogue.

Unless the dialogue is restored, there will be other, even more serious, breakdowns.

What’s your philosophy? – results

Table 1 shows the percentage of poll respondents who considered each item to be a real thing, along with the percentage who did not and those who were not sure. Respondents who did not reply make up the remaining proportion, the total adding up in each case to 100. Respondents were also asked to say what they thought the average professional physicist (table 2) and the average citizen (table 3) would think. A total of 534 replies were received.

1. What physicists think

	Real	Not real	Not sure	No reply
The Earth	93	3	2	2
Stones	93	3	2	2
Colours	50	36	9	3
Wavelengths	72	19	5	4
After-images	42	32	21	5
Hallucinations	40	43	13	4
Emotions	49	20	26	5
Genes	83	8	4	5
Atoms	84	7	5	4
The Bohr atom	18	68	10	4
Excited states of atoms	77	14	4	5
Electrons	84	9	3	4
Quarks	68	11	16	7
Higher-order infinities	26	32	37	5
Light waves	68	20	7	5
Viscosity	66	23	7	4
Kinetic energy	67	23	5	5
Electrical resistance	70	20	5	5
Mass	76	8	11	5
G (gravitational constant)	49	34	13	4
Real numbers	66	26	3	5
Imaginary numbers	43	44	9	4
The Ptolemaic solar system	9	70	16	5
The Copernican solar system	43	43	9	5
Wave-function (state of system)	42	43	10	5
Direction of time	43	38	15	4

2. What physicists think the average professional physicist thinks

	Real	Not real	Not sure	No reply
The Earth	93	0	1	6
Stones	93	1	1	5
Colours	45	37	13	5
Wavelengths	73	14	7	6
After-images	32	37	26	5
Hallucinations	24	54	16	6
Emotions	42	34	20	4
Genes	88	5	2	5
Atoms	88	3	5	4
The Bohr atom	20	57	18	7
Excited states of atoms	81	9	4	6
Electrons	86	3	6	5
Quarks	84	4	7	5
Higher-order infinities	32	49	15	4
Light waves	74	9	12	5
Viscosity	72	10	14	4
Kinetic energy	73	10	13	4
Electrical resistance	73	8	14	5
Mass	83	7	6	4
G (gravitational constant)	69	14	13	4
Real numbers	64	19	12	5
Imaginary numbers	63	20	12	5
The Ptolemaic solar system	6	71	13	10
The Copernican solar system	50	26	20	4
Wave-function (state of system)	50	19	27	4
Direction of time	51	18	26	5

3. What physicists think the average citizen thinks

	Real	Not real	Not sure	No reply
The Earth	94	0	1	5
Stones	93	0	0	7
Colours	81	6	8	5
Wavelengths	34	21	39	6
After-images	26	32	37	5
Hallucinations	29	47	19	5
Emotions	75	10	10	5
Genes	64	22	10	4
Atoms	82	2	12	4
The Bohr atom	21	10	64	5
Excited states of atoms	19	12	63	6
Electrons	67	4	24	5
Quarks	31	21	43	5
Higher-order infinities	13	32	49	6
Light waves	60	10	34	6
Viscosity	60	11	25	4
Kinetic energy	57	13	26	4
Electrical resistance	56	9	30	5
Mass	84	7	4	5
G (gravitational constant)	27	19	50	4
Real numbers	69	11	15	5
Imaginary numbers	8	68	19	5
The Ptolemaic solar system	7	73	16	4
The Copernican solar system	43	6	46	5
Wave-function (state of system)	12	41	42	5
Direction of time	69	8	18	7