Words can leave an impression when we read a perfume description before smelling it. But even when a description does not explicitly say for men or for women, could the gender information embedded in its wording affect how we remember the scent?
A study published by Laura J. Speed and Asifa Majid in 2019 focused on grammatical gender in French and German. The researchers paired commercial fragrances with experimental ingredient descriptions and tested whether participants later recognized the perfumes when smelling them again.
Grammatical gender is different from associations evoked by words
In French and German, nouns are assigned to grammatical categories such as masculine and feminine. This is grammatical gender. It does not mean that the objects themselves are masculine or feminine, and words referring to the same object can belong to different categories in different languages.
For example, the three nouns pumpkin, sage, and marjoram used in this experiment are masculine in German and feminine in French. The researchers created descriptions that conveyed the same content but differed in grammatical gender.
Separately, words may evoke associations with masculinity or femininity. Such associations relate to society, culture, and experience, and need not match grammatical categories. The study also considered whether a perfume itself smelled as though it were intended for men or women. These are three distinct kinds of information.
The first experiment compared two languages
Experiment 1 included 30 native German speakers and 31 native French speakers. All knew multiple languages; 41 were women and 20 were men.
Participants learned eight fragrances, including Dior’s J’adore and Chanel’s Bleu de Chanel. The fragrances were sprayed onto plastic pellets placed inside opaque squeeze bottles, which were squeezed to present the odors. The experiment did not show participants branded bottles.
Before each fragrance, participants read a description consisting of three ingredient names. The nouns within a description shared the same grammatical gender, and articles explicitly marking gender were omitted. These were experimental descriptions listing plausible ingredients, not descriptions matched to the actual formulas. Which fragrance was paired with which grammatical gender varied across the lists assigned to participants.
After smelling and rating the fragrances, participants were presented with the original eight mixed with eight new ones and indicated whether they had smelled each before. The task was to recognize the odors themselves, not to answer by recalling perfume names or descriptions.
The analysis reported a pattern of better memory when the grammatical gender of the description matched the fragrance’s perceived gender. However, the central statistical result was borderline at p = .05. When masculine and feminine fragrances were considered separately, the effect of the description’s grammatical gender was not significant. The congruence effect remained in another analysis, but the results do not establish a reliable memory improvement for every fragrance.
The second experiment manipulated associations separately
Experiment 2 included 42 women who were native German speakers. This time, the researchers combined grammatical gender with the gender associations evoked by words. For example, peach, honey, and lavender are masculine nouns in the German used in the experiment, but were selected as words with feminine associations based on ratings from a separate group of English speakers.
The procedure again used eight fragrances and eight new ones. However, one product selected as a men’s fragrance in the first experiment had been rated as feminine in a separate assessment and was replaced. Commercial categories and impressions from smelling were not necessarily the same.
In this experiment, there was a tendency toward better memory when grammatical gender and the words’ gender associations did not match. However, the difference had a p value of .051 and did not meet the conventional 5% criterion. It was also nonsignificant in another analysis that more fully accounted for differences among participants and fragrances.
Thus, the advantage of congruence in Experiment 1 was not simply replicated. The combinations of information being compared were different, too. Neither matching the description to the perfume’s impression nor deliberately making them inconsistent can yet be selected as a practical rule for improving memory.
Not only memory, but also how clearly ingredients were perceived
In Experiment 2, ratings of how clearly participants could smell the described ingredients were higher when grammatical gender and the words’ gender associations matched. Congruence with the fragrance’s perceived gender was also relevant. This was not a test of whether participants correctly identified actual ingredients; a description seeming plausible may have been linked to its ingredients seeming clearer.
Descriptions with feminine associations also increased stated willingness to buy the fragrance for a mother or sister, while those with masculine associations increased stated willingness to buy it for a father or brother. These were questionnaire responses, not observed purchases. Neither experiment found significant effects of the descriptions on pleasantness, intensity, or the amount participants were willing to pay.
Could words change the cues used to remember a scent?
Participants did not report realizing that grammatical gender or similar features were being manipulated. The authors emphasize the possibility that subtle linguistic information can enter fragrance cognition even without consciously using male or female labels.
To explain the different memory patterns across experiments, the authors suggest that multiple kinds of gender information within the same words might interfere, or that inconsistent information might remain as separate cues. However, the memory difference in the second experiment was not conclusive, and these mechanisms were not established. The study also did not distinguish a change during initial encoding of the fragrance from a change in later recognition decisions.
Summary
Perfume descriptions may do more than communicate ingredient names. This study suggested possible relationships of grammatical gender and social associations with memory and the perceived clarity of described ingredients. At the same time, memory findings differed across experiments and analyses, and there was no uniform change in pleasantness or intensity.
The important contribution is the expansion of research on language effects beyond overt promotional wording to linguistic features that usually pass unnoticed. Participants were predominantly women who knew multiple languages and used languages with grammatical gender for nouns. Rather than directly transferring the results to Japanese descriptions or real shopping behavior, future research needs to establish how fragrance complexity, language experience, and word associations interact.
The research was supported by an NWO VICI grant from the Netherlands Organisation for Scientific Research.
References
Speed, L. J., & Majid, A. (2019). Linguistic features of fragrances: The role of grammatical gender and gender associations. Attention, Perception, & Psychophysics, 81, 2063–2077. DOI: 10.3758/s13414-019-01729-0 / Full article. © 2019 The Author(s). CC BY 4.0.
