Friday, May 13, 2022

Smellosophy


The world of smell goes unnoticed to most of us. There are more books in the Library of Congress on perfume bottles than there are about smell itself. The year 2020 will mean lots of things to lots of people, but most of us will not remember it as the year that changed the literary world of smell forever.  

Image credit: Camillo Golgi's image of a dog’s olfactory bulb from his Sulla fina anatomia degli organi centrali del sistema nervoso, 1885. [link]

Together, Smellosophy by Ann Sophie Barwich and Nose Dive by Harold McGee have advanced the written record of our sense of smell by 40 years. Trygg Engen wrote The Perception of Odors in 1982, and we haven't seen a real update since. Obviously there have been scores of researchers writing journal articles since then, but an article is not a book. Also, there have been books written about smell, by well-regarded scientists and writers like Avery Gilbert, Rachel Herz, Synnott, Classen and Howes, Wilson and Stevenson, Alain Corbin, to name a few that come to mind.

But these books, both of them, are something else. They are dense, they are exhaustive, and they offer the most detailed explanation of how your sense of smell works, and what things smell like, than any other book you could read about the topic. 

*For third place in the most important smell-books of the past 40 years, maybe we should add The Essence: Discovering the World of Scent, Perfume and Fragrance by artbook publisher Gestalten, also in the year 2020/2021, and runner-up to Nose-Dive for The Perfumed Plume's best fragrance book of 2021

I'll save Harold McGee for another day, since I'm still parsing his almost 700 page codex. I thought it more important to finish Barwich's book first; it's the more scientifically hardcore of the two, and I was afraid my attention might wane the other way around. And this is not a book you want to breeze through. Again, it's probably the most comprehensive book on smell ever written. I took notes, some of which I'll paste here:

Notes on the author: Sophie Ann Barwich is a a cognitive scientist and empirical philosopher with a doctorate in odor classification and a background in philosophy and history, and spent time in Stuart Firestein's Columbia lab, and also interviewed a ton of perfumers, fragrance industry professionals, and just about anyone else who's important in the olfactory world that's still alive (except Asifa Majid, although she is cited in the book). She's an Assistant Professor at Indiana University Bloomington, between the Department of History & Philosophy of Science and the Cognitive Science Program.

Your nose is tailored to measure the world as calibrated by your mental life and physiological conditions. (p12)

Important terms for the olfactory enthusiast: Combinatorial and combinatorics refers to the combination of massive datapoints into a single datapoint, like how the olfactory bulb takes thousands of chemicals and pulses out a single signal in response; Foregrounding like from this sentence, "Smell is frequently embedded in the conscious experience of the world without being foregrounded as an olfactory experience" (p91); juxtaglomerular cells means next-to-glomerulus cells, juxta-anything sounds interesting, so...

Referencing Asifa Majid: Odor language is strongly contingent upon the rules of conventionalization (p102)

  • The Lingua Anosmia is Perishable: Leslie Vosshall - One central problem with these new computational studies were the data; "Most of the theoretical work has been based on a single 30 yr old dataset. Why has no one done an update?" (p173)
  • Andrew Dravnieks Atlas of Odor Character Profiles: "a great list in the early 80's, for use in the Northeast of the United States, for people who are baby boomers." (p173)
  • But so many of the words on that list have no frame of reference for the people who come to our studies. Any of these lists...they are perishable, highly culturally biased lists, that will work for some specific period in history, for a specific target audience. (p173)
  • The Odor Atlas didn't map odor quality space, "they mapped the odor quality space of Dravnieks." (p173)
  • The DREAM Project did provide a strong case for data mining, but only had a 0.3 correlation. (p175)

Polar surface area is a key metric for olfactory receptors (p185) -Poivet et al. "Functional Odor Classification through a Medical Chemistry Approach", Science Advances 4 n2 (2018)

Terry Acree's potato chips, only 3 odorants do the trick: methanethiol (rotten cabbage), methionol (potato), 2-ethyl-3,5-dimethylpyrazine (toast) -Computing Odor Images. Rochelle MM, Pr√©vost GJ, Acree TE. J Agric Food Chem. 2018 Mar 14;66(10):2219-2225. doi: 10.1021/acs.jafc.6b05573. https://pubmed.ncbi.nlm.nih.gov/28285523/

Stimulus Representation Beyond the Map: Gordon Shephard and Thomas Cleland: Olfactory perception is more like the feature coding of face recognition in the visual system. The olfactory bulb does not represent chemical classes but the chemical environment; it tracks the statistics of a changing odor environment (Shephard and Cleland, p233) [And I would say that society and culture are part of that environment, and that given enough data, the language of smell can be a map of our changing social environment.]

"The brain evolved from the body and not the other way around" -Terry Acree (p237)

Predictive powers of the nose, Walter Freeman studying neural networks at Berkeley in the 1980's:
How brains make chaos in order to make sense of the world. Skarda, C. A., & Freeman, W. J. (1987).  Behavioral and Brain Sciences, 10(2), 161–195. https://doi.org/10.1017/S0140525X00047336. https://psycnet.apa.org/doi/10.1017/S0140525X00047336
Simulation of chaotic EEG patterns with a dynamic model of the olfactory system, Walter J. Freeman, Biological Cybernetics (2004) v56 p139-150. https://link.springer.com/article/10.1007/BF00317988
Model of biological pattern recognition with spatially chaotic dynamics, Yong Yao and Walter J. Freeman, Neural Networks (1990), v3 p153-170. https://www.sciencedirect.com/science/article/abs/pii/089360809090086Z
Neural networks and chaos. Freeman WJ, J Theor Biol. 1994 Nov 7;171(1):13-8. doi: 10.1006/jtbi.1994.1207. https://pubmed.ncbi.nlm.nih.gov/7844992/
Characterization of state transitions in spatially distributed, chaotic, nonlinear, dynamical systems in cerebral cortex. Freeman, W.J. Integrative Physiological and Behavioral Science 29, 294–306 (1994). https://doi.org/10.1007/BF02691333. https://link.springer.com/article/10.1007/BF02691333

Unknown Odors (Covid) - based on Walter Freeman's work: "known odors elicit an established spatiotemporal signature of activity. Unknown odors first evoke chaotic activity before acquiring their own spatiotemporal signature for future recall. Chaos here was a condition for learners so that the brain would not confuse a novel odor with the signature of an already known one." (p238-239)

"The spatiotemporal activity in the bulb should thus be seen as an expression of the dynamic coding space -- not a fixed representation of odors, since odorants can be assigned various meanings, and, in turn, patterns." (p242)

The wide distribution of decorrelated signals in the olfactory cortex "allows olfactory signals to be integrated and synchronized with parallel processes in neighboring cortical domains..." Sparse coding is less detailed but faster in processing and recognition. Temporal patterns is where it's at, not topographical patterns, so it's about measurement, not mapping. "dynamically encoded signatures" (p242-243)

"Rather than molecules, your brain depicts transient information patterns, extracted and weighed in a given context, without a superimposed matrix of chemical classes to accommodate for countless permutations." (p246)

"Olfaction becomes an ideal model for higher-order processing other modalities.

Higher-brain integration is notoriously tricky to understand; it's signaling is not topographic, seemingly random, and autoassociate, just like in the olfactory system." (p247)

Relearning to Smell, Post-Covid: Mark Stopfer: "The first time you present an odor, there are no oscillations, you have to present an odor two or three times before the oscillation begins to build up. That's because there's this activity-dependent plasticity that takes place within the antennal lobe [we're talking insects here]. The local neurons that are activated become more effective over repeated activations. The inhibitory local neurons become more and more effective at synchronizing the projecting neurons over the course of repeated odor presentations. We think that's enabling the system to become more specific as the odor remains present." (p257)

"You go from a very general response to a more specific response. At the very beginning, you get this big burst that tells you there's something novel in the environment. Then right after that, you start to categorize it: it smells floral versus savory, for example. If the odor is still present, the system becomes more and more specific as this process builds up, the response downstream becomes more specific, and then you can identify exactly what it is. The same circuit at first will give you this generalization: it's something [fruity] -- and then the same circuit over time will say: oh well, it's cherry, not strawberry. It only happens if the odor is there long enough to perhaps be of interest to the organism." [this is why we can't name bad smells]

Beautiful description: The olfactory brain measures "odor situations" to evaluate how cues are related to each other (temporally, combinatorially, causally) and to attribute these perceptions a specific value (pleasant, putrid) and behavioral response." (p260)

Odor images are not encoded in the stimulus; they constitute mental impressions that arise from the categorization of sensory information. (p268)

One of a Kind: "When we manufactured a standard solution of an odorant, all we have to do is make another one, and it will smell different every time." -Terry Acree (p269)

  • Christophe Laudamiel's description of what a perfumer does - refinement of observation through cognitive engagement: "We don't have a super nose, but we notice things; it's our job to recognize a lot of smells, we pay attention, we recognize what we smell. We know how to describe things, and we know how to compose." (p272)
  • Christophe again: Perfumery is more than the sum of its parts: "What about black olive? when you want black olive, it's burnt rubber with wood." (p286)
  • Christophe again: Odorants convey more than one qualitative note: "I don't know a single molecule. You say cut grass? Cut grass is a whole world. In cut grass is wet dirt. There is a pear note. There is a green note, which you would say is a green, leafy note. But then how do you define a green, leafy note? That's the one that smells like cut grass. So it's a catch 22." (p297)

Visuocentric theories often follow the idea that perception is all about the stable representation fo objects [but odors are constantly changing.] (p303)

The brain is dynamic; it measures the world rather than mapping it. (p304)

Individual variation is not at odds with the notion of objectivity in perception; rather it is an expression of the core mechanisms of sensory systems ... The traditional dualism between objectivity and subjectivity in sensory perception presents itself as an artifact of older philosophical framing. It is time to change... (p311)

Notes:
Smellosophy: What the Nose Tells the Mind
Ann Sophie Barwich, Harvard University Press, 2020

Nose Dive: A Field Guide to the World's Smells
Harold McGee, Penguin, 2020

Personal criticisms:
If you're not already familiar with the science of olfaction, this book is not an easy read. That's mostly understandable, because it is so comprehensive, airtight in fact. Less excusable is another pattern I noticed, which is that it's hard to follow the quotations and remember who is saying what, and that's because people are referred to by their first names; this is a multidisciplinary crowd, from philosophers to chemists to perfumers to neuroscientists; it's very unlikely that the reader would be on a first name basis with all these people. Using their full names would give the reader a little bit more to hang onto in their working memory. For example, in another, completely unrelated book Chimpanzee Culture Wars, author Nicolas Langlitz continues to write linguist Michael Tommasello's full title after already having mentioned him like 300 times throughout the book.

Next, although less severe of a criticism, is that at one point the book changes in tone quite dramatically, almost as if it were two different books (circa p210). And last thing, which isn't a criticism but a simple note: She opens her chapter 9 with Parmesan Vomit (p264), which I called "Quantum Hedonics" in Hidden Scents in 2015.

No comments:

Post a Comment