Making Sense of Scents: The Science of Smell
Ann-Sophie Barwich
In 1914, Alexander Graham Bell proposed a challenge:
Did you ever try to measure smell? Can you tell whether one smell is just twice as strong as another? Can you measure the difference between one kind of smell and another? It is very obvious that we have very many different kinds of smell, all the way from the odour of violets and roses up to asafoetida. But until you can measure their likeness and differences you can have no science of odour.
This challenge remains unanswered today. However, it has been taken up.
Olfaction is a fairly young model system in sensory neuroscience. Before the discovery of the mammalian olfactory receptor genes by Linda Buck and Richard Axel in 1991, only a few dedicated scholars were trying to make sense of scents. For centuries, smell did not attract much attention from scholars in the sciences nor the humanities. One major reason for this neglect is suggested in in the opening quote by Bell. The sense of smell is incredibly hard to study in a controlled setting. For most of the history of science, olfaction has thus been set aside for its apparent lack of testability. While we have a better understanding of the olfactory pathway today, many of the central questions remain unresolved. How do you classify smells and how do you make their perception comparable? (And how do you control the volatile stimulus, its concentration, and its administration in psychophysical studies?) What are the perceptual dimensions of smell? Are there such things as primary odours? How does the brain represent smells? From this perspective, the discovery of how the sense of smell works presents us with an intriguing, yet untold, history of creativity in scientific reasoning. What is more, the crucial part of this history is contemporary!
Tony Alter, ‘A Nose for the Road’
My research is a current and cross-disciplinary analysis of olfaction as a new model system for the senses and in neuroscience. My focus is on scientific decision-making in laboratory practice. What kinds of epistemic and pragmatic factors influence the choice of concepts, the design of experimental strategies, and the interpretation of data in scientific practice? Must we consider different philosophical arguments about the advancement of science when we deal with ongoing research instead of historical case studies? Large parts of the philosophy of science have dealt with past science. Almost all central ideas in debates on such topics as model-choice, scientific progress, or the problem of underdetermination of theories by evidence have been developed and tested against historical case studies. Furthermore, support for our philosophical theories rests primarily on selected resources from the literature. The question that arises is how different our central concepts (such as empirical success) would be if we were to look at ongoing and present practice. What empirical, epistemic, and social factors did we miss that analysis of current practices can reveal, and that might shed further light on our studies?
I am interested in the conceptual foundations of (neuro)science, and I analyse the uncertainties and ‘missing knowledge’ that drive contemporary research. In the tradition of HPS, I also recover the hidden record of olfactory experiments in past science (from 1600 onwards) to analyse the creative strategies that a few dedicated scholars have used to make the sense of smell scientifically accessible. My approach towards integrating these historical ideas into current practice involves identifying older studies that are relevant to similar experimental set-ups today. One such example concerns the parallels I have found between the study of airflow patterns in the nasal cavity in nineteenth-century physiology and the modern recreation of these airflow patterns through 3D printing.
The study of smell provides an important opportunity to make philosophical and historical ideas central to experimental practice because almost all key players in the scientific development of the study of smell are actively conducting research. And this is where philosophical analysis meets science in a complementary fashion; the perceptual dimensions of smell are an open issue in both philosophy and science. For this reason, I work in close collaboration with the lab of Stuart Firestein at Columbia, and I am in contact with additional neuroscience and food-science laboratories at other US universities, in order to address three questions: How should we model the perceptual structure and the hedonic (affective) or aesthetic experience of smell? Is olfaction different to other sensory systems such as vision? And how are smells represented in the brain?
You may ask what characterizes smell as a model system for research on the senses. Recent studies in olfaction promise new insights into molecular biology and neuroscience. And these insights present a good incentive to rethink our traditionally visuo-centric approaches to perceptual analysis. Indeed, the difficulties that used to exacerbate the study of smell (and contributed to its historical neglect) have now turned into a matter of scientific interest. The content of odour perception does not seem straightforward. For example, unlike in visual images, we have difficulties with identifying the single constituents of smell mixtures (an older study by David Laing in 1989 showed that our ability to identify individual components decreases rapidly with mixture complexity). These and many other sensory puzzles spurred recent neuroscientific interest into the nature of smell, and such puzzles resonate with a variety of philosophical questions that have captured my attention.
However, too many misperceptions about smell remain popular. These misperceptions have fostered the long-standing lack of interest in olfaction, especially in the philosophy of mind. One was its persistent image as a ‘lower sense’ that lacks cognitive significance. (Thank you, Enlightenment.) This view has been corrected by the sciences, where many popular opinions about our sense of smell were exposed as misguided. For example, dogs are not necessarily better at tracking smells than humans. Quite the contrary: a now famous study by Noam Sobel tested thirty-two hungry Berkeley undergrads and showed that they can track chocolate in the same way that dogs sniff out pheasants. Our difficulty in assigning names and labelling odours is not a fact about human biology, but simply a result of (modern, Western) cultural neglect. (Of course, you don’t have to take my word for it. Just talk to fragrance chemists. They dazzle you with the richness of their odour language.) Moreover, it turns out that human olfaction is not in evolutionary decline either (as Pinker, Sagan, and many other popular science writers would have you believe). Instead, its significance shifted from orthonasal (inhaling) to retronasal (mouth-breathing) smelling, a process that is fundamental to our refined capacities in flavour perception. The importance of smell in tasting has been emphasized and recently brought to wider attention by neuroscientists such as Gordon Shepherd, as well as philosophers and cognitive scientists such as Barry Smith and Charles Spence.
Meanwhile, a general theory of smell and flavour perception has yet to be established. Flavour has significantly shaped human culture and history, and its multi-dimensional and cross-modal character invites a plethora of philosophical questions about perceptual content and dimensions. In answer to these questions, a better understanding of its neuroscientific basis and current scientific development is essential.
Figure 1
How do we smell, and how does our brain make sense of the scents in our nose? At first, the olfactory system seems like a straightforward three-level pathway (Figure 1). Odour perception begins when odorants (smelly molecules) are recognized by the olfactory receptors that are situated on the cilia of the olfactory nerves in the nasal epithelium. The receptor signals are collected in spherical neural spheres (so-called glomeruli) in the olfactory bulb at the frontal lobe of the brain (a domain that really looks like a bulb!). Notably, each sensory neuron expresses only one receptor gene (coding for one receptor type), and all neurons expressing a particular receptor are collected in one glomerulus. What you end up with here is a beautiful and neat activation pattern for each odorant—a fingerprint for each smelly molecule in the brain, if you will! This topographic organization of the bulb was expected to be maintained throughout further processing stages in the olfactory cortex. This is how it would work if this were the visual system, and one might reasonably expect it to be the same for olfaction. Now, however, it turns out that the higher-level processing of smells might be organized differently.
The open question is how the brain maps odours as a non-spatial, multidimensional stimulus. This points to an interesting difference with the visual system in its topographic representation of a spatial stimulus. As a cognitive scientist, I do wonder here: What are the implications of a ‘missing’ (or at least hidden) map in the olfactory cortex for our understanding of smell perception? One potential implication is that there may not be a central cortical domain where odour activation signals are synthesized into a unified smell percept. This brings us to something comparable to the so-called binding problem in vision: How does the brain compute a synthesized percept from different and segregated processing domains? To date, this question also remains open. In light of this, olfaction emerged as a new model system for understanding the principles of higher brain organization over the past two decades. Thus, for the philosophy of science, its current dynamics and susceptibility to the revision of its core premises makes olfactory research an excellent example for studying the ambiguity of determining what is a reliable research strategy.
The results of my research are currently being turned into a monograph about olfaction as an emerging model system in neuroscience and for theories of perception. This will address how changes in experimental and cognitive strategies define the meaning of central ideas in olfaction (e.g. olfactory primaries, perceptual dimensions of smell, the olfactory brain). The book will also include a transcribed discussion of key questions in olfactory research based on my interviews with several olfactory researchers. These edited dialogues provide a glimpse behind the scenes of ongoing research and the pluralism of scientific opinion about open questions. It further offers an insight into the kinds of philosophical questions and ideas of interest to practicing scientists. In fact, if we want philosophers of science to successfully exchange with practitioners, it is important that we move beyond the proclamation that has been repeated ad nauseam by some philosophers, namely, that scientists should listen to us. Scientists need not listen to philosophers—they should talk to us (and, that said, we should talk more to them).
Ann-Sophie Barwich
Columbia University
ab4221@columbia.edu