Complex odours and simple smells

I have just finished reading the introduction to a thesis. It was well-written and well-argued, a pleasure to read. Yet, what struck me was the fact that almost all of the literature referenced was published within the past 20 years. Now, for a student, 20 years ago may seem like ancient history, and of course there’s little point in going back too far if your field is molecular genetics. But for literature relevant to food or taste or odour preferences and perceptions is not like that: there’s a lot of psychology from mid-20th century (and sometimes earlier) that is highly relevant.

In 1968, Zajonc published the initial paper on the phenomenon of mere exposure (ME) [1], the idea that liking for something increases as we are repeatedly exposed to it. Notice that I didn’t say “as we become more familiar with it”, and this is because Zajonc found that, strangely, familiarity did not increase for his rapidly-presented visual stimuli even though liking did. This is just one of the many interesting aspects of this phenomenon. Indeed, his research participants weren’t even aware of the nature of the stimulus to which they were exposed since the exposure time was typically about 0.004 seconds.

Fast-forward forty years, and my colleagues and I studied ME for odours, which turned out to be different form the ME seen with simple visual stimuli [2]. We were interested in effects of attention on development of odour liking following studies that showed if you ignored a novel visual stimulus, it became less liked. We were familiar too with research showing that emotions direct attention – you’ll pay increased attention to something that is either appealing or threatening, for example. In fact, we managed to replicate the findings of the studies using visual stimuli, showing that you can be repeatedly exposed to an odour without getting to like it if your attention is directed elsewhere. It is as if, by allocating attention away from a stimulus, you are also sending a message to the emotion processing bits of your brain about its lack of emotional relevance.

Now, a recent paper by Delplanque and colleagues has shown that the ME effect with odours seems to depend on just how pleasant the odour is judged prior to exposure [3]. Importantly, the effect is not linear – that is, if an odour is either very pleasant or very unpleasant to begin with, you get little impact of being repeatedly exposed to it. This is not difficult to explain with very pleasant odours as they may be reaching a sort of hedonic asymptote. Also, since these are odours with which we are likely to be familiar, they may be more easily identified – in essence, we’ve already made up our mind about how pleasant they are.

Initially very unpleasant odours remain unpleasant despite exposure quite probably because their “hedonic tone” is a signal of potential danger – a signal that would be diluted if ME could change liking for them. As an example, odours of rotting, faecal odours, or highly pungent odours may provide us with good adaptive information because of their ability to repel us.

So, odours can be seen to be quite different to meaningless visual stimuli, of the type used in Zajonc’s studies, because unless they are highly novel, there is already some degree of pleasantness built in. Moreover, complete odour novelty is hard to find since the notes in complex odours may be reminiscent of other, better known smells. It is telling, for example, that the odour that showed the greatest change in the Delplanque study was the allegedly “no odour” control.

Familiarity is well known to be associated with increasing pleasantness in odours, but only if we exclude those odour that are highly unpleasant, for which the association breaks down. If for a moment, we do exclude these, does that mean that we like odours
because they are more familiar? Is this what ME is achieving for odours?

These questions bring me to the older paper that I really wanted to discuss, namely Berlyne’s studies of novelty, liking and stimulus complexity [4]. Like Zajonc, Berlyne used repeatedly presented visual stimuli and he initially found that lower familiarity was associated with high pleasantness. He realized, though, that he was failing to consider the complexity of that stimulus. To take account of this, he proposed that increased complexity – and/or increased novelty – produces high arousal, which is unpleasant. So, if we reduce the complexity and/or novelty of a stimulus, it becomes paired with lower arousal, and hence produces increased feelings of pleasure/reward.

So how does this apply to odours? Surely, we get bored with simple food or other odours, while complex odours are interesting? Well, maybe. But what constitutes complexity as far as odours are concerned? Let’s take the example of wines, mainly because we are used to talking about them in complex terms. It is evident that highly complex wines – those with many different facets or odour notes – are valued highly. You don’t get 95 on the “100 point” scale (see:
Overrated wines) by having a simple flavour. But ….. such points aren’t awarded for liking; they are awarded for appreciation, which is not the same thing at all. You and I can like a complex Burgundy an awful lot, even if the only term that we can apply to it is “wine flavour”.

In fact, and perhaps surprisingly, as we become more familiar with odours, we not only like them more, we also judge them as less complex [5]. But interestingly, familiarity is not associated with the number of notes that an odour might have – something that we might ordinarily think of (as in the case of wine, above) as a proxy for complexity.

So, attention, arousal, novelty, liking and complexity are all linked somehow, and if your brain is hurting, I can’t blame you – it’s all too complex! Try this. Zajonc proposed that exposed stimuli become like because of increased ease with which the brain can process them. Complexity may be synonymous with novelty simply because the brain takes longer or expends more effort in this processing. In terms of perception and behavior, something that is novel or surprising is, by definition, a violation of expectations that will involve increased (unpleasant) arousal and a direction of attention to stimulus features to assess if some action needs to be taken – perhaps there is danger, for example.

Consider this sequence next time you encounter something wrong or unexpected with, say, a drink. Arousal increases as you pay close attention to your drink: you sniff, repeatedly, attempting to identify the problem …. this process is, in a way, making features stand out and the drink odour seems more complex. You sniff again – hmmm, cat’s pee. That’s ok, then. You asked for a Riesling, but you were given a Sauvignon blanc. Arousal decreases, as familiarity and liking both increase, and you can pay attention to something else, like your meal.


1. Zajonc, R.B., Attitudinal effects of mere exposure. J. Pers. Soc. Psychol., 1968. 9(2): 1-27.
2. Prescott, J., H. Kim, and K.-O. Kim,
Cognitive mediation of hedonic changes to odors following exposure. Chemosens. Percept., 2008. 1: 2-8.
3. Delplanque, S., G. Coppin, L. Bloesch, I. Cayeux, and D. Sander,
The mere exposure effect depends on an odor's initial pleasantness. Frontiers in psychology, 2015. 6: 911.
4. Berlyne, D.E.,
Novelty, complexity, and hedonic value. Percept. Psychophys., 1970. 8: 279-286.
5. Sulmont, C., S. Issanchou, and E.P. Koster,
Selection of odorants for memory tests on the basis of familiarity, perceived complexity, pleasantness, similarity and identification. Chemical Senses, 2002. 27: 307-317.