Timeline of the metric expansion of space, where space, including hypothetical non-observable portions of the universe, is represented at each time by the circular sections. On the left, the dramatic expansion occurs in the inflationary epoch; and at the center, the expansion accelerates (artist's concept; neither time or size are to scale).

February 24, 2023, by Brigitte Nerlich

Can metaphors hinder scientific progress?

This is a guest post by Jack Morgan Jones. He is a postgraduate researcher at the University of Manchester’s Philosophy Department with an interest in truth and practical rationality, as well as agency and constructivism.


It’s readily acknowledged that metaphors can help an educated public better understand a scientist’s technical work. But questioning the role metaphors play in relation to a scientist’s own understanding touches on a slightly more sensitive topic.

Metaphors are not just tools for communicating with the broader public or educating in the classroom. Metaphors are also devices that scientists use to understand their own objects of interest. There are many different reasons a scientist might make use of such a device. A scientist might be dealing with a subject matter of immense complexity and only require an abridged understanding for use in their inferential reasoning. A scientist might require a simplified understanding of a subject matter adjacent to the domain they actually specialise in. Or, more controversially, a subject matter might defy human understanding without the help of metaphorical devices.

These metaphors are more common than we might think. For example, the image of the Big Bang represents the entire universe all the way from the initial singularity to the far-reaching ever-expanding edges of the universe. It’s a well-known image, but it has a subtle metaphorical structure that simplifies features of time and space to present the entire universe.

As you look at the featured image above, you are, as the observer of the image, excluded from being placed within the object represented. The very fact that there is a perspective observing the image from outside the image conflicts with the notion that the image is inclusive of all possible locations from which such a perspective could be taken.

Maybe this metaphorical representation does little work for a theoretical physicist. I suspect the importance of metaphor varies on a case-by-case basis, but I take it as a given that metaphors can help at least in principle. There are just too many examples that crop-up in scientific theorising, from the “landscapes” of developmental systems theory to the many metaphors (in their many states) that help with quantum mechanics. But can metaphors also obstruct scientific theorising?

From the history of science comes a warning call. Donna Haraway’s work on primatologists showed how their metaphors cast primates in the role of humans. 20th century primatology was awash with images of apes striking human-like poses or being held in human-like positions. As Haraway put it, these images served as “complex illusions” generating scientific facts: researchers spent their time selecting individual animals based on the idea of “typical specimens” and in so doing presupposed the things their study was meant to verify, such as whether human and animal aggression was alike in structure.

On Haraway’s reading, primatologists were held captive by an epistemological assumption that the world was structured in the image of man. As a result, questions concerning subject matters such as motherhood were predetermined, as was the functionalist lens through which they could be answered. Of course, similar unwarranted assumptions might have been shared by other research fields in the 20th century without bad metaphor playing this role. The risk though is that when bad metaphors strike these assumptions become more deeply embedded (the anthropological work of Emily Martin suggests a similar case of bad metaphor for US reproductive biologists of the 1970s).

Determining just how badly the aims of a research group become compromised can seem too hard an ask. It would require giving a counterfactual history: what would the field of primatology have achieved had it not been for a picture holding it captive? Answering such a question is tricky because there will be many historical confounders, and even those confounders that are not lost to history will be hard to assess relative to the significance of the bad metaphor.

Nevertheless, if the logic of the day screams out for the field to catch up, then we can hypothesise that bad metaphors play some such role. As Haraway pointed out, even in the 1980s primatology hadn’t yet evolved past a research agenda which had been blinkered by the idea of male-domination in the animal kingdom since the 1920s. The history of science warns us that scientific metaphors require care and attention.

Contemporary science will not be immune to the perils of metaphors. Its practitioners are still human, after all. Bad metaphors that get away from scientists can take some time to exhume. On the one hand, that they resist identification like this might make it overly ambitious to identify contemporary cases of bad metaphor, much less gauge their significance. On the other hand, what is the point of learning the lessons of history if one doesn’t try to put them into practice? Scientists should try out these speculations.

What might today’s bad metaphors be? Where might they be lurking? The most recent case I can gesture towards is the ‘genes as blueprints’ metaphor. The metaphor implies that individual genes correspond with different behavioural traits. The problem with this metaphor isn’t that the map isn’t the territory; it’s that the map might turn out to be poked through with holes and covered with ink-splotches! The metaphor is a major oversimplification given that genes are responsive to the environment they encounter. The assumption that genes play this one-to-one role might have been hindering scientific research, justifying research agendas that presuppose hard natural limits between groups, and selling the wares of commercial market peddlers.

A firm hand is required lest our metaphors run riot. I’ll finish with an illustrative story. The physicist Arthur Eddington once claimed that although there were solid objects such as tables, science had revealed that every table was in fact made of particles, and so tables had no real solidity. There were, Eddington claimed, two tables – it was just that we could only perceive with the naked-eye the one that appeared to us as solid. Now this makes sense if you understand particles as solid billiard balls floating around in empty space at the atomic level – then you really would have a table made of almost nothing. But the philosopher Susan Stebbing pointed out Eddington’s mistake: he had been led into thinking that for something to be solid all of its parts must be solid. A bad metaphor had run riot! – what Eddington should have realised is that to be consistent either he should never lean on a table again, or else accept that tables were solid, that they could be leant on, and that whatever the billiard balls were that made up a table, there was no common-sense way in which they could be described as solid.

It seems like we should want a criterion that sorts the good metaphor from the bad. I wonder whether such a criterion wouldn’t itself be susceptible to bad metaphor, and whether the role that metaphors play can ever really be settled by identifying some necessary and sufficient conditions. The best that science might have to hope for is that its practitioners read widely: that they engage with the work of those who study what and how scientists talk and write, and with the diverse ways in which different people can arrive at an understanding of the same subject matter.

Image: Wikimedia commons (Timeline of the metric expansion of space, where space, including hypothetical non-observable portions of the universe, is represented at each time by the circular sections. On the left, the dramatic expansion occurs in the inflationary epoch; and at the center, the expansion accelerates (artist’s concept; neither time or size are to scale).)




Posted in LanguageMetaphors