Defining Science: Language, Method, and Objectives
Nick Spencer
Read as a PDF.
Introduction
The British have never felt a particularly pressing need formally to define science. In 1843, Parliament passed the Scientific Societies Act, which exempted from the payment of rates “societies established exclusively for purposes of science, literature or the fine arts.” The Act has remained on the statute book ever since but in spite of its age, there hasn’t been a single case in a British law that has sought to define “science” for the purposes of the Act.
Not so America. The resurgence of anti-evolutionary views there from the 1960s was accompanied by a lexical shift, in which the anti-evolutionary movement, its institutions, publications and arguments, appropriated the language and framing of science. Biblical literalism was now presented through the endeavours of the Institute for Creation Research, and the Creation Science Research Center, and the Creation Research Society. The success in animating public opinion resulted in attempts to put ‘creation science’ on curricula, ideally on a parity with the teaching of evolution.
The consequence was, inevitably, a series of court cases from the early 1980s in which first creation science and then its child Intelligent Design were judged (and found wanting). The process of this generated various definitions of science through the courts, such as the famous Overton ruling from Arkansas in 1982, which stated that that the “essential characteristics” of science were:
(1) It is guided by natural law;
(2) It has to be explanatory by reference to natural law;
(3) It is testable against the empirical world;
(4) Its conclusions are tentative, i.e., are not necessarily the final word; and
(5) It is falsifiable
Other court cases – such as Edwards v Aguillard and Kitzmiller v. Dover Area School District subsequently reference and elaborate on these criteria, sometimes in judgements more often in submissions.
As it happens, the UK seems to be heading in a similar legal direction now, although for less alarming reasons. In 2006, the Charities Act was passed which included science as a charitable activity and this prompted the Science Council to clarify what the word ‘science’ meant. Following a year-long consultation, the Council proposed a definition of science as “the pursuit and application of knowledge and understanding of the natural and social world following a systematic methodology based on evidence.”
Now, neither the Overton ruling nor the Science Council are wrong in their definition. I suspect that any definition of science from anyone qualified to offer it would come up with something similar. But these examples do still invite a few questions.
The first is simply, who says? By what authority does the Arkansas court or the Science Council adjudicate on the meaning of science. The fact that both court and Council arrived at their decisions on the basis of submissions and consultations strongly suggests that the authority resides not so much in the court or the Council itself but in those whom they consult. Put another way, the meaning of the word in question resides in the way it is understood and used by those qualified to do so.
This should be recognisable as coming close to an idea that Wittgenstein developed in his Philosophical Investigations. In contrast with the earlier ‘atomic’ and representational theory of language in Tractatus Logico-Philosophicus, in which the meaning of a word is to be determined by its depiction of facts about or objects in the world, Wittgenstein in Philosophical Investigations advocates an understanding of language which locates meaning in usage. The meaning of a word lies not in ostensive definition. Rather, it lies in the (often very different) ways people use the word. As Wittgenstein says pithily, for most if not all words, “the meaning… is its use in the language.”
This naturally worries people who prefer clarity, conciseness and objectivity in their definitions. Doesn’t such an approach leave the door open to hopeless levels of vagueness. Arguably it does but then again if we are talking about broad, complex, socially-mediated phenomena like science, we should be prepared to accept a degree of imprecision.
Indeed, that is precisely what the Science Council does for, after its helpfully concise definition, the Council went on to elaborate on it. Or more precisely, it went on to elaborate what the pivotal word “methodology” in the definition incorporated. Their list was:
1. Objective observation: measurement and data (possibly although not necessarily using mathematics as a tool)
2. Evidence
3. Experiment and/or observation as benchmarks for testing hypotheses
4. Induction: reasoning to establish general rules or conclusions drawn from facts or examples
5. Repetition
6. Critical analysis
7. Verification and testing: critical exposure to scrutiny, peer review and assessment.
This is an impressive list but – note – its various articles are neither necessary nor exclusive. Some things that we might call science have problems with ‘experimentation’ or ‘verification’. Other things we wouldn’t normally call science, such as journalism or police work, or indeed certain humanities, are replete with ‘observation’ and ‘induction’.
There is certainly a relationship here, but it is an unclear one. And this points to a second way in which Wittgenstein’s ideas on language can help us.
After his description of various games in Philosophical Investigations, Wittgenstein concludes by remarking that the games he has discussed comprise “a complicated network of similarities overlapping and criss-crossing,” and, in the following section, he says the best way he can find to characterize this pattern of similarities and dissimilarities is in the idea of “family resemblances”.
By this analogy, members of a family will resemble one another in terms of their “build, features, colour of eyes, gait, temperament, etc. etc.” No-one will exactly resemble another family member, and no-one will share no common features. At the same time, there is no single feature that all family members possess which could be judged to be the essential and ubiquitous family feature. Identity does not reside in any single feature or, in the case of science, any one single disaggregated dimension or constituent element.
This means that it is impossible to say definitively whether something is (in Wittgenstein’s illustration) a game, a point that he readily acknowledges when he asks himself “What still counts as a game and what no longer does? Can you give the boundary?” and answers simply ‘No.’ A ‘family resemblance’ approach to language precludes making definitive definitions. It avoids essentialism.
This seems to be precisely what we get to and what we need when it comes to science: a cluster of features that between them, in some number and some combination, form a resemblance that we call science. It’s a messy approach – without the simplicity or elegance of a single allegedly authoritative dictionary definition – but it’s an approach borne of the fact that science is a messy, complex, shifting, socially-grounded and contested enterprise.
It’s worth noting that this is also the approach adopted by Massimo Pigliucci and Maarten Boudry in their edited volume on reconsidering the demarcation problem, 30 years after Larry Laudan’s famous essay marking ‘The Demise of the Demarcation Problem’. Pigliucci advocated what he termed “the possibility of understanding science as a Wittgenstein-type cluster concept.” Rather than demanding necessary and sufficient conditions, as had Laudan, science was better understood in terms of “fuzzy [logic] and gradual distinctions. Furthermore, it’s an approach that, as we shall see, has the potential to defuse a number of generally unnecessary conflicts that hover round science.
Introducing the Research
At this point I want to introduce the research project on which I have been working for a number of years. Between 2019 and 2022, I have been co-PI on research project into public and elite understanding of science and religion, which was funded by the Templeton Research Trust. The project comprised two elements of research.
The first was a qualitative study of over a hundred in-depth interviews with the leading thinkers and public communicators in the fields of science and of religion (and in the field of science and religion) in the UK. This involved interviewing a wide range of experts – including philosophers and sociologists of science and of religion; people who work in astronomy, astrophysics and physics; evolution, biology and genetics; mathematics, chemistry and engineering; psychology, sociology and anthropology; religious studies and theology, as well as those with a particular cross-disciplinary interest in science and religion. The second element of the research, which I won’t be drawing on in this essay, was a quantitative study of a nationally-representative sample of over 5,000 UK adults, assessing their understanding and opinion of science, religion, and science and religion.
The qualitative research allowed us to put to our experts a series of questions around what they thought science was, before, crucially, probing their answers in depth. In this way, we were able to explore language – specifically and narrowly the word ‘science’ – as it is used, and to do so in such a way that enabled us to fill out the full picture of the family resemblance. What did we find?
The first thing to say was that while all of our ‘elite’ interviewees were able to answer the question of what they thought science was, and many did so at length and in considerable detail, not many did so willingly. A large proportion, from the outset, expressed significant reservations about doing so, and many insisted that it was an inherently artificial and unsatisfactory exercise.
“Even just saying ‘science’ is problematic… There are various sciences [but] I don’t think there is such a thing as science… Science is not reducible to a single idea, there are lots of different sciences… I don’t think science is a natural kind or has a core essence… I don’t think there’s a scientific definition of what science is. ”
Interviewees recognized the imprecision of any boundaries put around science. Others remarked on their porosity and their mutability. Physics was frequently cited as the paradigmatic science and yet as a paradigm, its methods (and results) were judged to be “too distinct” from those of many other ‘scientific’ disciplines to be helpful in defining the overall category of science.
In a similar way, just as there were recognized sciences that appeared to fail to meet some fundamental scientific criteria, there were numerous comparable counterexamples of disciplines or “enquiries” that were not usually labelled science, such as history, journalism or detective work, but which did meet certain ‘scientific’ criteria pertaining to “methodology, truth seeking and so forth”.
Interviewees were much preoccupied with this liminal space of disciplines that were contestably scientific, citing as examples linguistics, social science, sociology, economics, behavioral science, political science, meta-cosmology, history, literature, psychology, anthropology, mathematics, archaeology, philosophy and formal logic, sports science, engineering, ethnographic research, biblical studies, and psychiatry. One brave soul (and not even a theologian) even made a case for theology being included in the category!
When pushed, they did develop their definitions from which we drew six factors or dimensions or, if you prefer to stick with the family metaphor, ‘features’ of science. There pertained to
1. The methods of science
2. The subject of science
3. The presupposition of science
4. The objectives of science
5. The values of science, and
6. The institutional structure of science
The Methods of Science
The idea that science could be defined and distinguished by ‘the scientific method’ was the most widely discussed definitional element within the interviews. However, at the same time, the idea that there was such a thing as ‘the scientific method’ was treated in much the same way as the idea that there was such a thing as science. Defining the scientific method and planting the flag of science foursquare upon it was, it was argued, futile.
Rather, there were elements or methods within the practice of science – or more precisely the theoretical practice of science (we will come to this important point later) that were characteristics of science.
First, the methods of science involve forming “questions”, “hypotheses”, “conceptual models” and/or “theories”about ‘reality’ (a word deliberately left opaque at this point). Second, science’s hypotheses (and theories) were testable and ideally refutable. Third, the testing of hypotheses and theories was grounded in a particular set of rules “practices” or “methods” that characterized science, namely the “ordered” observation of the “external” or “natural” world (though both those terms, as we shall see, were inherently contestable.) Fourth, commitment to experimentation was paralleled by commitment to quantification and accuracy. Indeed, the former was effectively irrelevant without the latter. Fifthly, this approach was provisional. People spoke sometimes of science being “provisional” and sometimes of it being “open”, “tentative”, “sceptical”, “adaptive”, “dynamic”, or epistemically humble. Between them these five sub-elements made up first feature of science – its methodology.
The Subject of Science
Second there was the subject of science. This has been a particular focus within US legal definitions of science, because of the particular tussle with ID, and it emerged clearly from our interviews too.
What science studies emerged as a significant dimension in the broader understanding of science, almost as important as, and inextricably tied up with the question of the way that science studied it. However, the variety of terms deployed in discussion of this issue (again) underlines how unstable and contestable this issue is.
We heard that science was a means of understanding “nature” or “the natural world”, or “the physical world”, or “the material world”, or more precisely “how the world functions at a materialistic level.” Alternatively, it was a way of understanding “the external world” or “the observable world” or the world to which we have “common access”. Precisely what constituted the category of the ‘natural’ was far from clear. At very least, it was historically contingent, with events or processes that once appeared non- (or even super) natural now being recognised as perfectly natural. There were similar questions around the idea of the material or the physical. Several interviewees, physicists in particular, raised questions about how material or physical their subject of study was.
None of these tergiversations around the words natural, material, physical, external, or observable amounted a wholesale rejection or undermining of the idea that science was properly speaking characterised by its subject as well as its method. On the contrary, people were clear that science did have a proper subject of study. It was just hard to pinpoint what that subject was.
The Presuppositions of Science
The extent to which science’s commitment to naturalism is itself an unprovable presupposition is highly contentious. For a philosopher like Alvin Plantinga, there is a presupposition to that effect, and it is a deeply problematic one. As he wrote in his book Where the conflict really lies, “there is a deep and serious conflict between naturalism and science”. By contrast, according to Judge Jones’ ruling in Kitzmiller v. Dover Area School District, “methodological naturalism is a ‘ground rule’ of science today”.
Interviewees did acknowledge that science was predicated on certain ‘presuppositions’ (again, precise terms varied here) and that these were fundamental, rather than incidental to its nature. First, there was a commitment to existence of specific laws and, more broadly, the immutable lawfulness of reality. Science was commonly described as “the process of trying to uncover laws” or “the systematic study of nature to extract laws”.
A second axiom lay in the conviction that the universe was intelligible or, put another way, that the evolved human brain was capable of grasping reality. According to this argument, science “believe[d]that the world is amenable to rational explanation.” (emphases added). A third, final presupposition was, so to speak, deeper than either of these, namely a prior commitment to the existence of the external world and to the reliability of human senses. As one philosopher remarked, “I don’t think there are any good philosophical arguments that it is okay to trust our senses, but most of us think…that all knowledge begins with trusting experience.”
These comments underline how science was also commonly characterised by having certain crucial and ultimately unprovable presuppositions without which it could not work.
The Objectives of Science
These three elements or dimensions or features of science – its methods, subject, and presuppositions – are all crucial elements within the overall ‘family’ definition we are accumulating, but risk giving the impression of science being a primarily theoretical activity, abstracted from anything that human beings actually do. This issue – “the actual practice of science” for want of a better phrase – was repeatedly raised by interviewees, practising scientists just as much as sociologists or philosophers of science.
“We continue to allow this kind of myth of science being out there on a cloud almost like a deity, in effect, rarefied beyond the capacity of human beings and science proclaiming on this and science proclaiming on the other. [But we should go] back to the original definition. Science is what scientists do.”
Accordingly, the second group of three dimensions or features – the objectives, values, and institutions of science – draw into the ‘family definition’ aspects of the empirical reality of science that are important but otherwise in danger of being overlooked.
The first of these – objectives – underlines how science is an attempt to achieve something. Science is characterised as “a reliable way of finding knowledge about the world”. That reliability was usually established by its ability to make accurate predictions about things. Science was (ideally) predictive. Science was characterised by its accurate predictions about reality and, moreover, because scientific knowledge was cumulative as well as predictive, those predictions got more accurate precisely because it was successfully predictive, and increasingly so over time, science was usually tied up with human endeavours to control, modify and improve the material world. It was not simply “a reliable way of finding knowledge about the world” but, as the same interviewee went on to say, something “that we can make predictions from, that we can use as the basis for technology and engineering and find that it works.”
Once again it is important to emphasise that within the family resemblance definition of science we are pursuing, not all characteristics are universally or equally evident in every area. Not all scientific disciplines could be equally characterised in this way regarding objectives. And even those interviewees who recognised science as having certain objectives did still draw a distinction between science and technology, or between science and engineering.
The Values of Science
Provisionality or epistemic humility was central to the method of science as outlined above. A number of interviewees spoke of this in ethical or cultural terms, as a value inherent in scientific practice. Science was not simply a process or a method. It was also “a way of looking at the world” , a “commitment to internal questioning and to critique”, “an attitude”, or “a self-critical aspect.” The importance of humility (also termed “intellectual… modest[y]”) as a value in science was incontestable.
Alongside humility, science was exemplified by curiosity. Indeed, in its own way, curiosity was as fundamental to science as any other factor so far discussed. In addition to humility and curiosity, there was an adherence to what might be termed self-discipline. This was present throughout the scientific process: in the commitment to objective (rather than subjective) explanations; to publicly-accessible (rather than privately owned or controlled) phenomena; to a broad, often collaborative and cumulative approach to acquiring knowledge (rather than narrow, personal intellectual advancement); and to subsequent application of that knowledge for a wider public (rather than private) good.
Beyond humility, curiosity and self-discipline, science was (or should be) characterised by a host of other virtues that play quiet and often unrecognised but nonetheless essential roles. There was creativity and imagination: “science depends on the imagination to create the hypotheses and decide what’s worth pursuing and all those sorts of things.” There was “intellectual integrity”: “moral commitments to truthfulness, to honesty… hard work… [and rejection of] negligence and deceitfulness”. And there was also instinct and intuition.
The Institutional Structure of Science
The sixth and final dimension to the family resemblance definition of science is its institutional nature. As with the values of science, this is well recognised in the sociology of science if less prevalent in the popular mind. It is, however, crucial.
At the first level, science is, as already noted, an inherently collaborative and communal enterprise, “a social activity”. So significant is this, that some claimed that the ‘solitary scientist’ was an oxymoron. Such collaboration has been thoroughly institutionalised. The communality of endeavour and attainment is not casual or informal but officially structured and recognised.
This institutionalisation manifests itself in (at least) two ways. The first concerns the specific institutions or organisations within which scientists work. This was typically, but by no means solely, universities. The second is the sense of belonging to a wider community of scientists, which institutionalises people through norms and practices around things like papers, peer reviews, journals, and conferences.
The institutionalisation manifests itself particularly with regards to language and structures of authority. Regarding language: in theory, science should open: “accessible in principle to third parties”, “a completely democratised route to knowledge”, committed to “public accessibility”. In reality, its institutionalisation gives its a linguistic identity which is often inaccessible or incomprehensible to those outside.
With regard to structures of authority: if as one sociologist put it, “science is what scientists do”, the following question – who, then, is a scientist? – can only be answered with reference to institutions. Scientists are people who are formally recognised as scientists. Science is “an institutionalised way of knowing”. The “order, method, and structure” of science “are agreed on by a community of knowledgeable peers.” “Science is a profession” and thus in some way dependant on formalised structures of professional recognition.
Conclusion
This messy list of features will disappoint anyone seeking a clear and well-delineated definition of science. However, science is too complex, diverse, heterogeneous, social-mediated, and fluid an activity to lend itself to a single, simple definition, or even a set of definitions.
This is not to say that all these features should carry equal weight. It is arguable that the cluster of features that gather round the methodological dimension of science are probably the most significant ones to note when it comes to thinking about how we define science. They are, if you like, the signature feature. However, recognising the significance of one particular feature should not blind us to the range of others that also play a role in delineating science.
The resulting, composite, “fuzzy” picture also has the merit of defusing various conflicts – between science and religion, or science and the humanities, or science and the arts, or science and politics – that a single, hard-edged definition risks.
We can draw a parallel here with economist Amartya Sen’s ideas about identity. In a number of books – most obviously in his collection of essays Identity and Violence – Sen argues that awareness of our multiple and combined identities as human beings can help mitigate ethnic and religious hatred.
If we understand cultures as single, discrete, and self-contained entities, especially ones in which we invest all our emotional and psychological energy, the chances of us being drawn into conflict to ‘defend’ them is far higher than if we understand cultures as multiple, overlapping, protean, and contestable. In this regard, Sen sets out a vision that places himself diametrically opposed to Samuel Huntingdon, whose famous thesis reified entire civilisations and the predicted a clash between them.
As I have argued elsewhere, if you reify something as complex and variegated as a civilisation, of course you are likely to see a clash around it. And perhaps we might say the same thing about science. If we see science as singular, monochromatic and hard-edged, we are more likely to find it in sustained and uncompromising conflict with other similarly singular and hard-edged entities. If, however, we understand it as a series overlapping family resemblances, the potential for absolute conflict is abated.
To be clear, conflict may still be present. You may, to take just one example, be convinced that a feature of science is its commitment to not only methodological naturalism but to full-on ontological naturalism, and then hold that as being completely incompatible with any religion worthy of the name. In other words, you may, like Alvin Plantinga, see the conflict as really lying between religion and naturalism, but unlike Alvin Plantinga, identify science as being necessarily naturalistic in the fullest sense, thereby making the conflict ultimately between religion and science.
Even if this is your view, however, it doesn’t mean you will totalise the conflict, as you may well see perfect compatibility between science’s methods, presuppositions, and values, and those of religion. In effect, a family resemblance approach to science offers a nuanced and empirically-justified understanding of what science is but it also offers a way of defusing some the more unnecessary tensions that sometimes surround it.
Nick Spencer is Senior Fellow at Theos, host of the podcast Reading our Times, and the author of Playing God: science, religion and the future of humanity. The Landscapes of Science and Religion: what are we disagreeing about will be published by Oxford University Press in 2025.
