Introduction
References to the connection between language and its affects on perception dates back to Plato’s Cratylus dialog. Clearly, many more experiments have been conducted and empirical data collected concerning the topic since that time. This paper will explore the linguistic influences on one specific aspect of perception: food. That is, it will discuss how and if language, defined as the sound of the name of the food, can influence one’s perception of how desirable or flavorful the food is. Such effects can be separated into three categories: the semantic, phonetic, and associative influences of gastronomical words on the gustatory sense. Unlike other commercial products, food engages at least three of the five sense simultaneously. Thus, due to this interaction of at the very least, the visual, gustatory, and olfactory senses, the addition of the possible influences of the auditory sense greatly complicates research, thus making direct experiments of this simultaneous relationship few, and empirical data scarce. However, experimentally extrapolated evidence for each of the three categories (semantic, phonetic, and associative) on their effects on human perception, though not necessarily in the directly gastronomical arena are abundant. Therefore, we will discuss such information in respect to the food industry. Additionally, numerous resources have pointed to the significant effect of the phonetic effects of words on one’s perception, and thus, following a brief discussion of possible influences incurred by semantic and associative affects, this present study will discuss in detail the influence of phonetic effects by citing both previous and original research conducted in attempts to make a connection between phonetic symbolism and the desirability of different unfamiliar foods. Specifically, we will explore what effects the auditory sounds of the branding of certain foods has on the perception of appeal.
Semantic effects
The semantic effect of labels on food can be discussed with a particular emphasis on the framing effect, in which different but logically equivalent words or phrases causes individuals to alter their decisions (Druckman, 2001). In respect to the food industry, this can be very well seen in a recent empirical example. In the past decade, the food industry has been increasingly projecting a more youthful image. Thus, language has been manipulated to fit this projection, and especially in the prune industry, has wrought significant benefits. Due to the association of prunes with torpid senility, the population in general consciously avoids such personal associations, and thus for the last 10 years, prune sales have been continuously falling. Logically then, in June of 2000, the California Prune Board lobbied the FDA to change the official name describing the desiccated fruit from prune to dried plume, a name which consumers claim to evoke a positive image of the beneficial affects of the fruit. Such a seemingly inconsequential change in the nominal aspect of a food item yielded inspiring results for the industry. After the label makeover was complete, the now dried plum industry saw an increase of 5% in sales. Thus, this example shows that language can be manipulated to increase appeal, even though every aspect of the core of the product remained unchanged. The connotative associations made with each respective label played a significant role in determining gustatory appeal. Another example of framing effects in the food industry is the switch of the naming of the Chinese Gooseberry to the now internationally recognized Kiwifruit. In these cases, it is apparent that language, through its semantic properties, can significantly alter consumer’s preferences and decision-making.
Sound-based effects
The second category of food names under consideration is the phonetic aspect of the gastronomical cognomen. Specifically, we explore the phenomenal of phonetic symbolism, which is defined as the “nonarbitrary relation between sound and meaning.” The explanation of phonetic symbolism has been divided into several categories. The first, vowels and consonants, explores the different impacts certain vowels and consonants have on perception. Vowels are usually separated by a front versus back distinction, which describes how the vowel sound is produced by the tongue. An epitomic example of a front vowels is the [i] in mil, and of a back vowel the [a] in mal. One experiment using these two vowels is described by the following excerpt:
“In what appears to be one of the first controlled demonstration of this effect, Sapir (1929) gave participants nonsense (artificial) words in the form of consonant-vowel-consonant that differed only in the middle vowel (e.g., mil vs. mal). He then gave participants an arbitrary referent (“these are tables”) and asked them to
indicate which was large and which was small. The participants showed over 80% agreement across a large number of word pairs in their association of the back vowel sound (e.g., mal) with a large table and the front vowel sound (e.g., mil) with a small table.” (Shrum and Lowry, 2006)
Such results were not only observable in English speakers, but also in Chinese, Navajo, and a variety of other languages, regardless of subject age or level of education. This indicates a potential universality of phonetic symbolism. The implications of these results have been used in fields such as advertising, when producers are in the process of naming a product. Particularly in the advertisement segment of the food industry, the front versus back distinction have informed firms that ice cream names with the [o] sound in frost are perceived to be creamier, smoother, and heavier than those names with the [i] sound in fish. The study that discovered this entailed providing a group of participants with two fictitious ice cream brand names, Frish and Frosh, and asked the participants to select which brand of ice cream appeals to them more. Enough of the participants chose Frosh to indicate that participants (and consumers, as a result) make an implicit connection between the creamy, smooth, and heavy connotations of the [o] sound and the desirable traits in ice cream.
Similar findings have occurred with consonants, although instead of the front and back categorization, phonetic symbolism effects of consonants are separated into fricatives or stops. The former, fricatives (i.e. s, f, and z), point to the sounds formed when air flows through the lips, teeth, and tongue, which causes friction, while the latter, stops (i.e. p, k, t, b), indicate sounds formed by the complete closure of aforementioned articulators. Additionally, consonants can also be separated into voiced and voiceless sounds, the distinction made by whether the vocal cords vibrate during sound production. The implications these categories have on human perception of descriptions will be discussed in further detail in relation to the original experiment.
Although such a relationship between sound and perception has long been observed, cognitive reasons as to why sounds of words cue words’ meanings are still debatable. Some researchers attribute it to synesthesia, which refers to a cross-modal sensory association. In this case, the stimulation of the brain by an auditory effect causes a gustatory perception. The cognitive specifics of this condition are largely debatable, but several possible explanations that have arisen include the discussion of a disinhibition of feedback and an excess amount of anatomical connections. However, the nuances of the degrees to which one sense can trigger another are still undetermined.
Associative effects
A combination of the semantic and phonetic effects of language on perception is the associative effect. Most commonly seen in advertising, this effect describes the usage of particular words that evoke desirable memories and/or perceptions that are automatically connected to the product in question, even if all they share is an auditory similarity. Few experiments have been conducted to explore the relationship between associative affects of food perception, but in the field or advertising, researchers have found significant reasons to believe that associative effects (called the language contact phenomena) are conspicuously present in product perception and selection. In one study, the researcher found that when introducing several European languages to Japanese advertising, certain consistent characteristics often linked the language of the advertisement and the consumer’s perception of the product advertised. For example, products with English slogans or brand names usually evoked a sense of “international appreciation, reliability, high quality, confidence, practical use, [and] practical life style,” (Haarmann, 1989) This is not a singular occurrence, for when conducting similar research but in the American car name industry, researchers found that “French names were also employed to conjure up connotations of fashion, elegance, and femininity, while Spanish was associated with freedom, adventure, and masculinity” (Piller, 1990). In another study by Kelly-Holmes, the link between food perception and the associative effects of language was briefly discussed in the finding that internationally, the Italian language connotes the perception of “good food and a positive attitude toward life.” Thus, any label that phonetically resembles of these foreign languages will evoke impressions that links the product to the qualities of said language.
Even though the above studies have sufficiently proved the presence of a correlation between language and perception, little research has been conducted to explicitly explore such correlations between language and food perception. Unlike in the cases of product advertisement, food perception brings in a third sensory faculty to the table: gustation. Thus, the following study has been created and conducted to explore the weight language carries on food perception, and seeks to ask whether under equal and simultaneous circumstances, auditory or visual influences dominate in the judgment of food appeal.
Present study: Empirical questions and predictions
The following study aims to further explore the correlation between phonetic symbolism and its effects on perception of different foods. It asks if there is a relationship between how much euphonic (front vowels) and dysphonic (back vowels) vowel sounds (and to a marginal extent, consonants) affect how appealing a certain dish is to the participant, and whether the pleasantness of sounds affect whether the participants perceive the dish to be sweet or savory, heavy tasting or light tasting, and whether they envision it to be served hot or cold. My predictions are that to an extent, independent of the appeal the visual image provides, words containing euphonic sounds will subconsciously be considered when considering these questions, and will yield results that indicate dishes with these sound will be more appealing, and sweeter, lighter-tasting, and served cold. Dysphonic words will seem less appealing, and imagined to be saltier, heavier-tasting, and served hot. Such hypothesis is based on past data, which showed that vowel sounds made with the back of the tongue (i.e. [oo]) tend to convey large, dull/less appealing, and more cumbersome features, while sounds formed by the tip of the tongue (i.e. [i]) had a lighter, sharper, and more stimulating connotation.
Method
Design
With the aims of this experiment in mind, a within person study was chosen was chosen to show the possible disparity in influences between euphonic and dysphonic vowel and consonant sounds, although with a major emphasis on vowel sounds due to there being fewer of them, and thereby facilitating and expediting the process and interpretation of the experiment. A within person study was chosen instead of a between study because the former produces stronger results and eliminates or accounts for several variables. For example if a between study was chosen instead, in which half of the sample group was shown pictures with euphonic names and the other half with dysphonic names, then effects of variables like gender, ethnic background, individual preferences, etc, would have to be considered, making for a very inefficient study.
Participants
Due to the nature of this study, there was no particular requirement for participants. The sample group that was tested was composed of all Stanford undergraduates residing in Otero House in Wilbur Hall, ranging from first to third years. Although 60% of the participants were female, this does not significantly influence the results yielded. No distinction was given to different ethnic backgrounds, although the sample represented a fairly wide spectrum.
Materials
This study required minimal supplies. All supplies needed were incorporated in the design of the experiment. As the researcher, I created a six-page powerpoint slide show each with a picture of a foreign, unfamiliar plate of food. To minimize externalities, I ensured that the photos selected from online sources were of similar sizes, clarity, and lighting. Therefore, I tried to minimize any difference in perception that could have been caused by different presentations of food, brightness of photo, and dominant color. All photos chosen were only moderately distinguishable (either covered in a sauce or had no definite shape or form), just enough to be recognized as being edible, but not enough to create a strong associative effect. Photos represented foods from the entire taste spectrum, ranging from sweet to savory to spicy, etc. Six artificial words were used, half of which had the front vowel sound [ee], created by the tip of the tongue, and the other half composed of the back vowel sounds of [oo] and [ä]. Artificial words were chosen to avoid any cultural or semantic associations. The words were listed in the following order: Noofura, Minizie, Teebigee, Slopuga, Grawkaw, Limiera. The name assigned to each picture was decided by a random computer sequence generator program to avoid any structural bias. The order of the pictures remained constant throughout the experiment, but the names changed with each subject according to the pattern generated by the program. The only other supply required was six identical survey slips (for each of the slides) containing all the same questions for each participant.
Procedure
After indiscriminately selecting a participant, a quiet setting was chosen to conduct the experiment, one where no one else was present to influence the decision-making process. Participants were all tested at about the same time of the day to avoid extremities associated with disparities in hunger-levels. I introduced the study with a cover story that avoided imparting any preconceived biases or influence. The participant was told that he/she was about to be taught some gastronomical vocabulary in the language of a foreign tribe (specific name was deliberately omitted to eliminate associative effects), and that throughout the test, he would be answering questions pertaining to how appealing the food of the tribe was to him. The participant was then told that for each slide, he would see a picture of the food, listen to me pronounce the name aloud once, and repeat it out loud three times before he took the four-question survey for each picture (sequence of questions: appeal*, sweet or salty, heavy or light tasting, and served hot or cold). The time limit for each survey slip was 15 seconds. The questions the participant had to answer are as follows:
* a four-point rather than a five-point scale was chosen to avoid the impulse to chose the moderate, middle-of-the-road rating of three when deciding under time pressure.
Discussion
The general yield of this test can be seen as having inconclusive results. I observed several experimental errors that affected the outcome. First, one of the food items chosen too closely resembled a brownie, something most if not all the participants were familiar with and instantly associated it with a positive appeal. Therefore, the results for this item consistently showed that participants imagined that they would like it, that it was sweet in taste, and would be heavy-tasting, indicating the preconceived notions associated with a typical brownie. Another uncertainty that could have caused a significant error margin is that the words could have been constructed with a pattern in mind. Specifically, the division between the front and back vowel sound words could have only been marked by a single letter change in the word, instead of utilizing an entirely different set of consonants and construction pattern.
However, there are certain results that suggest a possible correlation between sounds and food perception. With the previous example, only the cold or hot question did not yield a predictable result. Because everyone’s experience with brownies in terms of the temperature they were served varied, the results had much more potential to be influenced by the sound of the word assigned than the other questions. For that specific visual, the results indicated that slightly more than half of those tested accorded with my initial hypothesis. When this food was named with a word with the back vowel sound, it was reported as being served hot, and when it was labeled with a front vowel sound, participants reported that it was served cold. In fact, generally speaking, this particular question yielded the most conclusive results. Half of the selection of six showed that phonetic symbolism could have played more potent of a role in perception decisions. 54% of the participants reported answers that accorded with my hypothesis: foods labeled with the [oo] and [ä] sounds were envisioned as being served hot, while those with the [ee] sound were seen to be cold.
Another test that yielded intriguing results is the test of “heaviness” of taste. Because previous tests in brand name products suggested the association of front vowel sounds with lighter impressions and back vowel sounds with heavier impressions, this test was conducted to see if such results would also occur with the perception of food. The results were surprisingly conclusive, for regardless of which picture was labeled with which category of names, 60% of participants associated pictures with back sounds to be “heavy-tasting” and front sounds to be “light-tasting.” The results for the temperature and heaviness of flavor lie in accordance with Klink’s experiment that showed that front vowels convey attributes coldness and lightness in commercial products and back vowels with hotness and heaviness.
There doesn’t seem to be a significant relationship between phonetic symbolism and the perception of sweet or salty judgments, with almost all of the participants rating foods as either sweet or salty based on the visual impact, regardless of euphonic or dysphonic sounds. The same result occurred with the test of appeal, with judgments as to whether the subject would enjoy the item or not based solely on the impact of the photo rather than the sound of the name.
In conclusion, though some less tangible categories of perceiving foreign food can be moderately influenced by phonetic symbolism, visual imagery, and thus the sense of sight seem to be more dominant in the final decision. While running the experiment, I noted that many participants subconsciously remark that they rated items a certain way because it reminded them of a more familiar dish. This suggests the instinctive reaction of linking the unfamiliar with anything vaguely familiar, thus producing results that disprove my hypothesis. However, supportive remarks were also noted, especially concerning completely indiscernible and completely foreign items. For such items, several subjects remarked that they wouldn’t like it because the name sounded “disgusting.” This only occurred with items labeled with a [oo] vowel sound, thereby staying in accordance with Otto Jespersen’s note “that back vowels such as the [u] sound in dull or ugh are very often found in words expressing disgust or dislike” and shown empirically by Grant W. Smith in his paper, “The Political Impact of Name Sounds,” when he hypothesized that “if names of candidates contained vowel sounds that are often used to express disgust (e.g., putrid, puke), then candidates with last names containing such sounds (e.g., Dewey, Buchanan) might be less favorably perceived than other candidates with better‐sounding names” (Lowry and Shrum, 2007). In our case, it was food rather than political candidates that was perceived with more disgust. However, because this was merely a verbal observation and was not significantly reflected in the written results, we cannot conclude that euphonic sounds produce a more favorable perception of food. Therefore, this test suggests that the name of foods has a rather insignificant effect on appeal perception and only becomes even a consideration when considering the subtler details of the food presented, and is otherwise clearly dominated by visual images. However, the case for phonetic symbolism remains strong, and more potent tests need be conducted to determine the extent of the influence sound symbolism has on gustatory appeal.
Bibliography
Druckman, James N. Using Credible Advice to Overcome Framing Effects. University of Minnesota. Oxford UP, 2001. 7 Nov.-Dec. 2007
Hubbard, Edward M., and V.s. Ramachandran. Neurocognitive Mechanisms of Synesthesia.
Center for Brain and Cognition, University of California. San Dieto, 2005. 7 Dec. 2007
Lowry, Tina M., and L.j. Shrum. Phonetic Symbolism and Brand Name Preference.
University of Texas At San Antonio. JOURNAL OF CONSUMER RESEARCH,, 2007. 7 Dec. 2007
Piller, Ingrid. Advertising as a Site of Language Contact. Cambridge UP, 2003. 7 Dec. 2007
"Prune Gets $10 Million Makeover -- as Dried Plum." CNN. 13 Sept. 2000. 7 Dec. 2007 http://archives.cnn.com/2000/FOOD/news/09/13/prunes.reut/
Shrum, L.j., and Tina M. Lowry. Sounds Convey Meaning: the Implications.
University of Texas At San Antonio. Psycholinguistic Phenomena in Marketing Communications, 2007. 7 Dec. 2007 http://faculty.business.utsa.edu/tlowrey/LowreyBookChapter2006.Final.pdf
Yorkston, Eric, and Geeta Menton. Names on Consumer Judgments.
Journal of Consumer Research, 2004. 7 Dec. 2007
Zasky, Jason. "Turning Over a New Leaf." Failure Magazine. Nov.-Dec. 2007
