A still life painting featuring a white jug, an orange sphere, and two books against a deep blue background. The objects are arranged on a yellow-wooden and dark surface, with the orange resting on the book. The geometric composition is bathed in shadow and light.

Still Life with White Jar, Orange and Book (1932-33) by Vilhelm Lundstrøm. Courtesy the National Gallery of Denmark, Copenhagen


Philosophy is an art

For Margaret Macdonald, philosophical theories are akin to stories, meant to enlarge certain aspects of human life

by Peter West + BIO

Still Life with White Jar, Orange and Book (1932-33) by Vilhelm Lundstrøm. Courtesy the National Gallery of Denmark, Copenhagen

‘Philosophical theories are much more like good stories than scientific explanations.’ This provocative remark comes from the paper ‘Linguistic Philosophy and Perception’ (1953) by Margaret Macdonald. Macdonald was a figure at the institutional heart of British philosophy in the mid-20th century whose work, especially her views on the nature of philosophy itself, deserves to be better known.

Early proponents of the ‘analytic’ method in philosophy such as Bertrand Russell saw good philosophy as science-like and were dismissive of philosophy that was overly poetic or unscientific. Russell, for example, took issue with the French philosopher Henri Bergson, who was something of a bête-noire for early analytic philosophers. Bergson’s theorising (Russell thought) did not depend on argument but rather on expressing ‘truths’, so-called, arrived at by introspection. As Russell wrote in ‘The Philosophy of Bergson’ (1912):

His imaginative picture of the world, regarded as a poetic effort, is in the main not capable of either proof or disproof. Shakespeare says life’s but a walking shadow, Shelley says it is like a dome of many-colored glass, Bergson says it is a shell which burst into parts that are again shells. If you like Bergson’s image better, it is just as legitimate.

Russell places Bergson alongside William Shakespeare and Percy Bysshe Shelley and worries that there is no objective measure of whose worldview is more accurate. There’s no way of proving which is a better account of things, it’s simply a matter of which ‘image’ you like best. In other words, there’s no attempt to provide empirical evidence – evidence based on publicly observable data – in support of these views. For Russell, this was enough to show that what Bergson was doing was not really philosophy, at least not good philosophy, any more than Shakespeare’s plays and Shelley’s poetry were.

Russell’s view of what counts as good philosophy was not one that Macdonald shared. In her 1953 paper, she embraces comparisons between philosophy and literature, poetry and art. For Macdonald, philosophical theories are very much like ‘pictures’ or ‘stories’ and, perhaps even more controversially, she suggests that philosophical debates often come down to ‘temperamental differences’. For example, whether you are willing to believe (in accordance with thinkers like René Descartes) that we have an immaterial soul will come down to more than just the philosophical arguments you are presented with. Your view on this matter, Macdonald thinks, will more likely be determined by your own personal values, life experiences, religion and so on. In this way, she thinks, temperamental differences account for many philosophical disagreements.

However, unlike Bergson, Macdonald was not working in a different philosophical tradition from Russell. She was, to all appearances at least, just as much a part of analytic philosophy as he was. In fact, institutionally, she was at the very centre of things. Macdonald studied at the University of London and her PhD was supervised by Susan Stebbing, the first woman in Britain to be appointed a full professor of philosophy. Along with Stebbing and others including Gilbert Ryle, Macdonald helped found Analysisthe academic journal of analytic philosophy – which she later edited after the Second World War. And throughout the 1930s and ’50s, she published many articles in venues like the Proceedings of the Aristotelian Society, the UK’s foremost philosophical society, and was an active member of Cambridge University’s Moral Sciences Club.

So what happened? How did Macdonald end up with such a different view about what good philosophy looks like from Russell’s? And, if Macdonald was right, what does that imply about the value of philosophy?

The story of Macdonald’s entry into philosophy is quite remarkable. She was born in 1903 into poverty to a single mother who later absconded to Australia, leaving her baby behind to be fostered. Macdonald was ill throughout her childhood and youth, suffering from tuberculosis (among other things), and supported by an organisation called the National Children’s Home and Orphanage that later helped pay for her undergraduate studies. As Michael Kremer notes, Macdonald’s upbringing stands in stark contrast with many of the canonical figures in 20th-century philosophy such as Russell, who was born into British aristocracy (his grandfather was an earl who was twice prime minister), Ludwig Wittgenstein, a member of what was, historically, an extremely wealthy family (his father Karl, an industrialist, was one of the richest men in Europe), or Ryle, who spent his entire adult life easily moving through the ranks at the University of Oxford.

Stebbing was an important figure in Macdonald’s life, both personally and professionally. Macdonald was one of a number of women who benefitted from Stebbing’s supervision, along with Ruth Lydia Saw and Elsie Whetnall, and would go on to teach at Bedford College (now part of Royal Holloway, University of London) where Stebbing was professor of philosophy.

It was not easy for women to establish themselves in philosophy at this time and many women’s careers were negatively impacted by sexism (Oxford didn’t bestow degrees to women until 1920, and at the University of Cambridge it even later: 1948). When she applied for G E Moore’s Chair in Cambridge in 1938, Stebbing, for example, was told by Ryle that ‘everyone thinks you are the right person to succeed Moore, except that you are a woman’. Blunt to say the least. Similarly, in a letter to a friend in 1939, Macdonald writes: ‘I have been hoping to get a permanent lectureship in philosophy … It is difficult in my subject, especially for a woman.’ Nonetheless, through a pipeline from Bedford College into academic philosophy (and often back to Bedford College), Stebbing was able to help several women establish themselves in the profession.

She subjects philosophical enquiry itself to scrutiny, analysing the ways that philosophers talk and write

It is also likely that Stebbing had a hand in pushing Macdonald to focus on the relationship between philosophy and language. In Thinking to Some Purpose (1939), Stebbing emphasises the importance of distinguishing between different uses of language – eg, the difference between descriptive and emotive language – not only in philosophy, but in public discourse such as politics and journalism. After working with Stebbing, Macdonald went to Cambridge to work on a project on the influence of language on the concept of ‘matter’, a topic on which Stebbing herself once planned to write a book.

Soon after arriving in Cambridge in the 1930s, Macdonald met Wittgenstein, a towering figure in 20th-century philosophy. Along with Alice Ambrose, Macdonald attended many of his lectures – and the two women would go on to publish the so-called Blue and Brown Books, collated from notes they took between 1932-1935. It is no coincidence that, from the late 1930s onwards, Macdonald’s work often draws on what is known as ‘linguistic analysis’ – which is an approach to philosophy rather than a specific theory. Linguistic analysis was central to Wittgenstein’s later philosophy (and important in Ambrose’s philosophy, too).

Linguistic analysis involves paying attention to and drawing conclusions from the language used in particular contexts, including philosophical debates, scientific theories, and ordinary (common-sense) language. In her essay ‘Linguistic Philosophy and Perception’, Macdonald subjects philosophical enquiry itself to scrutiny, analysing the sorts of ways that philosophers talk and write – especially in comparison with scientists. This kind of linguistic analysis is a way of taking a step back and taking a look at the practice of philosophy itself. It involves answering questions like: What do philosophical disagreements involve and what do philosophical theories look like?

In a sense, then, Macdonald’s aims can be thought of as anthropological: she is interested in making observations about what a particular subsection of society – philosophers – are doing and providing a description of their activities. Macdonald takes philosophers of perception as her case study (hence the paper’s title) and, by paying attention to the language they use, offers an account of what philosophy of perception really amounts to. Although, as we will see, her findings stretch beyond just philosophy of perception, they include the nature of philosophy itself.

Putting the tools of linguistic analysis to work, Macdonald focuses her attention on the word ‘theory’. What do philosophers mean when they talk about philosophical ‘theories’? And is it the same thing that scientists mean when they use the word ‘theory’? Macdonald’s answer is a categorical ‘No’.

She claims that, when scientists put forward theories, they do so to explain empirical facts. Scientists put forward hypotheses (eg, ‘Earth is round’ or ‘physical objects are governed by laws of gravity’), which can then be verified (or falsified) by experiments and observations, leaving behind only plausible theories, and eliminating those that are refuted by factual evidence. Thus, Macdonald writes: ‘Confirmation and refutation by fact is an essential part of the meaning of “theory” in its empirical sense.’

If ‘confirmation and refutation by fact’ based on experiments is essential to the way that the word ‘theory’ is used by scientists, that provides a basis on which to examine whether philosophers use the word ‘theory’ in that way. And this is where Macdonald thinks philosophical theories differ from what scientists mean by the term:

They cannot be tested. Every philosophical theory of perception is compatible with all perceptual facts.

According to Macdonald, philosophical theories cannot be tested. Is that true? What might she mean by this? Once again, she uses the philosophy of perception as her example.

Philosophical theories, unlike scientific theories, are not in the business of discovering new facts

Two opposing positions in the philosophy of perception are direct realism and indirect realism (I’m going to oversimplify both here). Direct realism is the view that we directly perceive external objects in the world around us. When I look out of my window, I directly see a tree – and the nature of my perceptual experience informs me (directly) about the nature of the tree. Indirect realism, on the other hand, is the view that I only ever indirectly perceive objects like trees. What I directly perceive are mental representations – ie, ideas of trees – that are produced in my mind when my sense organs (eg, my eyes) are stimulated in the right way and send signals to my brain. I learn about the world around via these ideas (also known as ‘sense data’) in my mind. Direct realism might seem more common-sensical, but indirect realism might seem better equipped to deal with the existence of illusory or hallucinatory experiences, where I am seemingly not perceiving the world the way it really is. Given all this, isn’t it true to say that direct realists and indirect realists disagree on the facts?

In a sense, yes. But Macdonald’s point is that there is no disagreement on the phenomenological facts: facts about what it is like to have a perceptual experience. Both the direct realist and the indirect realist agree that, when I look out my window, I see a tree. What they disagree on is what it means to say that ‘I see a tree’ – they disagree on the mechanics of what is going on, or how best to explain the fact that I see a tree. Most importantly, for Macdonald, there’s no empirical test available to draw a line between the two theories. We can’t run an experiment to test for the truth of either theory because, on the level of experience, both parties agree that it’s true to say: ‘I see a tree.’

Thus, the first step in Macdonald’s meta-philosophical argument is to show that philosophical theories are not ‘theories’ in a scientific sense since they lack the essential criterion of being confirmed or refuted by fact. For this reason, she argues, philosophical theories, unlike scientific theories, are not in the business of discovering new facts.

So what is it that philosophical theories do? Macdonald’s answer is: ‘What they do suggest are new forms of expression for familiar facts.’

At this point, Macdonald’s analysis of the value of philosophy takes a turn that would have made Russell – who tried to move philosophy as far away from the arts as possible – very uncomfortable. Macdonald claims that philosophy’s value is much closer to that of art, literature or poetry than science. She explains that the arts inform us that ‘Language has many uses besides that of giving factual information or drawing deductive conclusions.’ A philosophical theory may not provide ‘information in a scientific sense’, she writes, ‘but, as poetry shows, it is far from worthless.’

By this point in her career, Macdonald had engaged extensively with the philosophy of art, the philosophy of art criticism, and the philosophy of fiction. Her meta-philosophical claims in ‘Linguistic Philosophy and Perception’ indicate that her engagement with the arts gave her an acute sense of where their value lies. What’s more, she evidently came to believe that the value of philosophy is very similar.

A good work of poetry, art or literature, Macdonald explains, can ‘enlarge’ certain aspects of human life to help us see and think about them differently. For example, Shakespeare’s Othello encourages us to think about jealousy by making it the centrepiece of the play. Or consider the emphasis on humanity’s relationship with nature in Romantic poetry. In both cases, the artist has ‘zoomed in’ on, or ‘enlarged’, an aspect of life – in a way that it is not typically enlarged in real life – to encourage the audience to reflect on it.

Hers is a 180-degree turn away from the ‘scientistic’ account of good philosophy Russell endorsed

Macdonald’s claim is that philosophical theories act in a similar way – different theories ‘enlarge’ certain aspects of experience. And this, in turn, means that the proponents of those theories end up telling competing stories. Some philosophers, like Plato, emphasise the degree to which our senses deceive us. Others, like Aristotle, emphasise the degree to which sense-experience is key to knowledge. Again, Plato and Aristotle did not disagree on the facts of experience – both agree that when I look out the window, I see a tree. But they disagree about the kind of story we ought to tell about those facts. In Plato’s story, the senses are the villains. In Aristotle’s, they are the heroes. Thus, Macdonald writes:

Everyone, it is sometimes said, is born either a little platonist or a little aristotelian. Whatever be the truth of this aphorism has little to do with the truth and falsity of these doctrines. It refers rather to temperamental differences.

What we find in Macdonald’s meta-philosophy, then, is a 180-degree turn away from the ‘scientistic’ account of good philosophy that Russell endorsed. Russell worried that choosing between Shakespeare, Shelley or Bergson might turn out to be simply a matter of individual preference, that there would be no criterion for showing that one was a better thinker than another. But Macdonald’s claim is that this is true of any philosophical theory – different stories will suit different temperaments.

At this point, one might think: enough is enough. It’s all very well to consider how philosophy overlaps with the arts, but surely Macdonald has gone too far when she suggests that philosophical theories are just ‘good stories’. More formally, one might worry that Macdonald’s account of philosophical debate generates a problem of relativism.

If philosophical debates come down to ‘temperamental differences’, then it looks like there’s no real right or wrong (or true or false) – any more than it’s right or wrong to prefer John Keats to Shelley, or Sally Rooney to James Joyce. Macdonald herself articulates the concern like so: ‘Ought not philosophy to be impersonal, unemotional and strictly rational?’

This leaves Macdonald with two options. The first is to bite the bullet and accept that, since artistic judgments are relativistic, and philosophy is like the arts, then philosophical preferences must be relativistic too. But there is another response available to Macdonald that does not involve accepting the charge of relativism. Note that the line of reasoning above depends upon a crucial assumption: that artistic judgments are relativistic.

Is this really true? Are judgments about art, literature and poetry purely a matter of subjective preferences? Some might be tempted to answer ‘Yes’. If I like my child’s hand painting more than a piece hanging in Tate Modern, I might be inclined to say that, for me, it is a better piece of art. Similarly, if I get more enjoyment reading Rooney’s novel Normal People than Joyce’s Ulysses, then who’s to say that Joyce is a better writer.

For Macdonald, the job of a moral philosopher is akin to that of an art critic

However, elsewhere in her writing – eg, her essay ‘Natural Rights’(1947) – Macdonald endorses the view that, while artistic judgments cannot be empirically tested – and thus ‘falsified’ or ‘verified’ like scientific hypotheses – they can be defended and justified. An art critic can justify their judgment that one piece of art is better than another. And they can persuade others to agree with them, and minds can be changed. In that sense, Macdonald suggests, a critic is like a barrister, pointing to certain evidence and telling a story intended to win over the jury to a particular point of view. In other words, artistic preferences are not entirely relativistic.

Consider a case in which you read a novel and find it underwhelming – it didn’t capture your imagination or engage you. But later you speak to a friend who explains how the novel alludes to certain literary tropes, or subverts the genre in some unique way, or satirises a political movement you were unaware of. You might find that your mind has changed. Your attention has been drawn to features of the novel and, to borrow Iris Murdoch’s words, you have been forced to ‘look again’.

In ‘Natural Rights’, Macdonald argues that ethical judgments (eg, ‘murder is wrong’ or ‘it is wrong to steal’), while they are not empirical in same way that scientific hypotheses are – they cannot be tested by experiment – are nonetheless meaningful. And, again, she draws on judgments about the arts as a model for what meaningful, but non-empirical, statements might look like. For Macdonald, the job of a moral philosopher is akin to that of an art critic: both are in the business of defending or justifying certain judgments or preferences. It’s not, as Russell says, as simple as liking one image more than another. There’s an onus on being able to justify or rationalise that preference.

The worry might persist that surely there’s the matter of truth to contend with. Philosophical theories might be like good stories, but surely only one of those stories can be true, or at least closer to the truth than another? Macdonald doesn’t address this question head-on so it isn’t obvious what her answer would be. I have attempted to show that her view is that even our individual preferences can be defended or justified, just like works of art, meaning that our philosophical views need not purely come down to mere gut intuitions. But I am tempted to suggest that Macdonald would not be overly concerned about truth – at least not in the way we usually think of it. Other scholars, like Cheryl Misak, have connected Macdonald to pragmatist philosophers like Frank Ramsey. Pragmatism, in a nutshell, is the view that what is true is what is useful. And different philosophical theories can be useful to different people for different reasons (such as, Macdonald might add, their temperament). While it is by no means explicit, the relativistic bent to Macdonald’s account of philosophical theories might signal that she was influenced by pragmatist ways of thinking about truth.

While Russell’s comparison between Bergson’s philosophy and the writings of Shakespeare or Shelley is intended as a form of criticism, Macdonald argues that an appreciation of the arts is key to understanding where the value of philosophical enquiry lies. In fact, Macdonald argues that philosophers ought to stop trying to make scientific philosophy a thing – because it is dangerous for philosophy. So long as philosophers like Russell keep up the pretence that philosophy ought to be like science, they are judging it by a standard that it cannot hope to meet – precisely because philosophical ‘theories’ aren’t empirically testable.

But Macdonald’s attempts to push philosophy away from science and towards the arts isn’t just a defensive manoeuvre. It’s also, she thinks, a way of making the value of philosophy clearer. For Macdonald, philosophy’s value lies not in providing us with new facts about the world, but rather in helping to see the familiar in a new light, in drawing attention to features of experience that might ordinarily pass us by, and by providing us with stories that can help make better sense of the world around us. Whether her story about what philosophy is is better than Russell’s story, or just a different story, well, that’s up to you to decide.