Cheese or Vomit? How Names Change the Way Odors Are Perceived

The same odor can receive very different pleasantness ratings depending on whether it is introduced as “Parmesan cheese” or “vomit.” This finding has been reported in actual olfactory experiments.

Does the influence of a name extend beyond how people answer a question? This article follows three studies: a 2001 experiment on changes in impressions produced by words, a 2005 brain-imaging study, and a 2008 investigation of sniffing and physiological responses. Although they share a theme, they used different odors and asked different experimental questions.

“Cheese” and “vomit” changed ratings of the same sample

In the 2001 study, 80 university students smelled the same five odors in two sessions one week apart. Each odor was given either a name with positive associations or one with negative associations. The names were switched in the second session, and their order was reversed between participant groups.

The researchers selected ambiguous odors for which more than one source name seemed plausible. For example, a mixture of isovaleric and butyric acids was labeled “Parmesan cheese” or “vomit,” violet leaf was labeled “fresh cucumber” or “mildew,” and patchouli was labeled “incense” or “musty basement.”

The largest pleasantness change occurred for the mixture labeled cheese or vomit. On a scale from 1, extremely unpleasant, to 9, extremely pleasant, the mean was 5.06 with the Parmesan cheese label and 1.68 with the vomit label. Without changing the sample itself, the rating shifted from approximately neutral to strongly unpleasant.

Besides rating pleasantness, familiarity, and intensity, participants freely described memories, uses, and what they themselves would call the odor. Two judges unaware of the experimental conditions classified these descriptions. For the acid mixture, approximately 83% in each presentation-order group were judged to have interpreted the two experiences as different odors. This was not the percentage choosing “different” in a direct same-or-different question.

The power of a name depends on the odor and presentation order

Labeling did not change all odors equally. The menthol labels, referring to breath-freshening mint and a chest medicine, produced a smaller change than the cheese example. When menthol was first introduced as medicine, its pleasantness increased with the subsequent mint label. In the group that received the mint label first, however, the pleasantness difference between the two labels was not clear.

Other ratings did not change uniformly either. The cheese/vomit mixture received lower familiarity ratings under the vomit label. Violet leaf was perceived as more intense when labeled mildew rather than cucumber. From these differences, the authors suggested that links between names and past experience help determine how words influence perception.

The authors proposed the phenomenon as evidence of olfactory illusions. Operationally, this means that context changes the interpretation of a physically unchanged stimulus. There was no unlabeled condition, and the odors were selected for ambiguity, so the experiment does not show that any odor can be transformed into anything simply by naming it.

Brain imaging compared “cheddar cheese” with “body odor”

The 2005 study used fMRI to investigate whether words alter brain responses involved in odor pleasantness. Participants were 12 healthy men aged 23–35. The sample combined isovaleric acid with cheddar cheese flavoring, so both the mixture and labels differed from the earlier Parmesan/vomit experiment.

The same odor was delivered by an apparatus for eight seconds while a screen displayed either “cheddar cheese” or “body odor.” Clean air was also given those two labels, with a separate clean-air condition labeled “air” serving as a baseline. Reference odors labeled as flowers and burned plastic were also used. Participants rated pleasantness and intensity after each presentation.

On a scale from −2 to +2, the sample’s mean pleasantness was −0.10 with the cheddar cheese label and −0.86 with the body odor label. Rather than becoming highly pleasant when called cheese, the sample shifted from approximately neutral to unpleasant when called body odor. Intensity ratings did not show a clear labeling difference.

Brain imaging showed a greater response to the cheese label than the body odor label in a region extending from the rostral anterior cingulate cortex to the medial orbitofrontal cortex, which is involved in pleasantness evaluation. Activity also correlated with pleasantness ratings. Responses associated with pleasantness or labeling were reported in the amygdala and neighboring olfactory regions when odor was present.

Interestingly, even clean air was rated more unpleasant under the body odor label than under the cheese label, and rating-related activity appeared in the anterior cingulate/medial orbitofrontal region. The amygdala, however, did not show the same clear association for labeled air. The authors discussed this as a distinction between regions influenced by word-induced expectations or imagery and regions more closely tied to actual odor input.

In a control test outside the scanner involving 11 people, no labeling effect was detected on inhalation during the first two seconds after stimulus onset. The authors interpreted the results as difficult to explain solely by changes in how much people sniffed and as evidence that semantic information contributes to olfactory processing. Brain imaging alone, however, cannot fully read a person’s inner experience or establish the causal pathway through a neural circuit.

With number labels as a baseline, pleasantness ranked positive, neutral, negative

A distinctive feature of the 2008 study was the use of neutral two-digit numbers alongside positive and negative labels. Forty university students smelled 15 odors, each under all three types of label. Additional samples were interspersed and presentation order was arranged to make repetitions less noticeable.

For example, citral received labels referring to squeezed lemon and insect repellent, while cumin oil was labeled as a curry spice mixture or dirty clothing. Participants read each name aloud, sniffed for two seconds, and then rated intensity, arousal, and pleasantness.

Across the 15 odors, pleasantness was highest with positive labels, followed by neutral numbers and then negative labels. Negative labels produced higher intensity ratings than positive or neutral labels, while positive labels produced greater arousal ratings than neutral labels. Individual odors did not all change in the same way, however: pyridine and cineole showed no clear pleasantness difference between positive and negative names.

Sniff volume and skin responses changed, but not in the same way

The same 2008 paper also reported a physiological experiment involving a separate group of 30 people. Using ten target odors and comparison samples, the researchers recorded inhalation, skin conductance, and heart rate as participants read names and smelled the samples. Because the participants differed from those in the rating experiment, the study did not directly test how closely an individual’s pleasantness changes matched that same person’s physiological changes.

When odor was present, sniff volume was greater with positive labels than with neutral or negative labels. After excluding poor recordings and other unusable data, 24 participants were included in the sniff-volume analysis. Peak sniff amplitude also showed a main effect of labeling, with positive labels exceeding neutral ones, but volume and peak amplitude did not show the same relationships with the experimental conditions.

Skin conductance responses were greater with both positive and negative labels than with neutral numbers. The difference between positive and negative labels was not significant, so this was not simply a pattern in which more pleasant meant a larger response. Heart rate showed no clear difference. The skin-conductance pattern remained after statistically adjusting for sniffing differences, leading the authors to suggest that changes in sniffing alone could not explain it.

When odorless water was presented, labels did not produce differences in skin conductance or sniff volume. However, participants had been told in advance that an odorless stimulus would follow the names in this condition. Expectations and outcome measures differed from those in the 2005 clean-air experiment, so the odorless-condition results cannot simply be treated as contradictory.

Summary

The three studies showed that odor names can change pleasantness ratings, with effects also appearing in brain responses, sniff volume, and some physiological responses. The Parmesan cheese/vomit experiment produced a large rating shift; the cheddar cheese/body odor imaging study found altered responses to the same sample; and the study adding neutral number labels found pleasantness ranked in positive, neutral, and negative order.

At the same time, labeling effects differed among odors, and physiological measures did not all move together. Our experience of an odor incorporates not only the compounds reaching the nose but also an interpretation of what the odor is. These studies explored that relationship using different methods.

The 2001 study received funding and odor materials from Haarmann & Reimer, and a Morley R. Kare Fellowship was listed for one author. The 2005 study was supported by the UK Medical Research Council, and three coauthors were affiliated with Firmenich at the time. The 2008 study received Canadian Institutes of Health Research funding. No separate conflict-of-interest statements were found in the original texts examined.

References

Herz RS, von Clef J. The influence of verbal labeling on the perception of odors: Evidence for olfactory illusions? Perception. 2001;30:381–391. doi:10.1068/p3179

de Araujo IE, Rolls ET, Velazco MI, Margot C, Cayeux I. Cognitive Modulation of Olfactory Processing. Neuron. 2005;46:671–679. doi:10.1016/j.neuron.2005.04.021

Djordjevic J, Lundstrom JN, Clément F, Boyle JA, Pouliot S, Jones-Gotman M. A Rose by Any Other Name: Would it Smell as Sweet? Journal of Neurophysiology. 2008;99:386–393. doi:10.1152/jn.00896.2007