Is there a relationship between sound and meaning?

Sound Symbolism in Mandarin Chinese

Note: I wrote this report as a part of a summer linguistics course at Columbia University during high school many moons ago. I’m reposting it here because it again is relevant to my life.

Introduction:

Have you ever wondered why some words sound like what they mean? A big bouncy bubble will bulge but a serpentine snake will slowly slide and slither. You may have thought of these words solely as onomatopoeia, words that sound similar to the sound they represent, but they are a lot more than just onomatopoeia. This phenomenon is known as sound symbolism, the “use of specific sounds or features of sounds in a partly systematic relation to meanings or categories of meaning” (Mathews, 1997).

I had always wondered why in many languages the word for mother started with the phoneme /m/ and the word for father started with a labial sound, ie. /f/ /p/ /b/.

I first became exposed to the phenomenon of sound symbolism upon reading about Vilayanur S. Ramachandran’s (2001) replication of an experiment by Wolfgang Köhler (1929, 1947) in which he discovered what is now known as the kiki/bouba effect. In this experiment, people were shown shapes similar to those in Figure 1 and asked which one they thought was kiki and which one they thought was bouba. 95% of the subjects picked the shape on the left as kiki and the one on the right as bouba. This was probably due to the sharp tongue inflection for the unvoiced /k/ sound, which makes it sound harder, as opposed to the softer voiced bilabial /b/ sound (Ramachandran & Hubbard, 2001). It is even thought by some that each sound comes with an inherent meaning. That is why so many words starting with /b/ have something to do with explosions or loud noises (Magnus, 2001). Especially after reading that language is a human instinct, I started to really believe the idea that certain sounds may cause us to instinctually attach some meaning (Pinker, 1994).

So profoundly amazed was I at this phenomenon, which I myself found to be true (something deep within me had told me to choose the spiky star as kiki, yet offered no explanation why), that I had wondered ever since: What other words had these properties? Does this work for all languages? Is there really a systematic relationship between sound and meaning?

And so my thoughts turned towards my own language, Chinese. Would Chinese languages have more sound symbolism than other languages? After all, it is the only language remaining that has words created out of pictograms, why should it not also have many more words based on sound? 

I had long noticed that the Mandarin word for heavy 重, pronounced “joong” or / ʈʂʊŋ/ with a falling tone sounded heavier than its counterpart 轻， pronounced “tsing” or /tɕʰiŋ/ with a high steady tone. This phenomenon is similar in Cantonese and no doubt many other varieties of Chinese, but due to my being most familiar with Mandarin, I decided solely to use Mandarin for this project. I was sure that this was no coincidence and so for my linguistics class at High School Programme 2007 at Columbia University, I decided to do my project on this topic.

Hypothesis:

My hypothesis at first was that yes, there is a relationship, but pretty soon, I realised that not only was that too simple a hypothesis, some words might not exhibit any relationship between sound and meaning at all. I categorised the words into concrete and abstract words, the former being nouns, verbs and adjectives that can be seen, and the latter being not visible. This, however does not include colours, although a study by Sabaneev and Pring (1929) showed a high 80% correlation between sound and colour (that study was done with musicians and I believe non-musicians will not relate sound to meaning). My hypothesis became the following: Concrete words exhibit much more sound symbolism than abstract words. This is because concrete words can be visualised whereas abstract words cannot.

Methodology:

I split the words into concrete and abstract and opted to omit parts of speech such as articles, conjunctions, prepositions and pronouns because I doubted any sort of relationship between sound and meaning for these words. I also omitted adverbs because these are generally similar if not identical to their adjective forms.

I would go out and survey at least ten people with words that can be seen in table 1.  These people would be anyone on the Columbia campus who seemed available and not busy. They would be asked if they had minimal of no prior exposure to the Chinese language. I assumed that subjects would take this to mean Mandarin Chinese as knowledge of other forms most likely are limited. If the subjects’ acknowledged minimal or no prior experience with the Chinese language, I would tell them a pair of definitions and a pair of Mandarin words and ask them to guess which one is which.

Results:

Overall, after tallying up the scores of the twenty subjects that were surveyed, I discovered that the number of correctly matched pairs slightly outnumbered the number of incorrectly matched pairs and that the concrete words were also slightly higher in number. There were 116 instances of correctly matched concrete word pairs and 64 instances of incorrectly matched ones whereas for abstract word pairs, there were 100 instances of correct pairs and 80 instances of incorrect pairs. The accuracy rates of concrete word pairs and abstract word pairs are therefore 64.4% and 58.8% respectively.

The next two pages are the results table of the survey I took and some notes I took down. The pronunciation is in what should be a quite faithful rendition of Mandarin Chinese phonology on the top, and on the bottom it is written in International Phonetic Alphabet symbols. The columns # correct and # incorrect refer to the number of instances that pairs were matched correctly and incorrectly

In Graph 1, the results have been broken down into its constituent categories of concrete and abstract nouns, verbs and adjectives and placed side by side for comparison. From this graph, we can see that while concrete nouns and adjectives generally beat their abstract counterparts by around 10% or so increased accuracy rate, in the case of concrete and abstract verbs, it was the abstract verbs that were more accurately identified by subjects by about 5%. 

Problems:

In retrospect, I have identified several areas of below-optimum performance of methodology. I have listed these below and have speculated the possible impact of these items. 

1. “Minimal prior exposure to Mandarin Chinese” may have still been enough for subjects to have learnt some of the easy vocabulary I had chosen. For example, the words for “name” (名字), “good” (好) and “love” (爱), are commonly taught in Chinese phrasebooks in sentences such as “what is your name?” “I’m fine, thanks” and “I love you”. The word “love” is especially susceptible to subjects’ prior knowledge as this word is identical in many cases in Japanese and to some extent in Korean. All three of these words were in the abstract category and subjects scored high on each of them, getting a range of 75 to 92 percent accuracy. 

2. As for the word pair “to love/to hate” (爱/恨), the high accuracy rate of 92% may also be attributed to the harsh consonant sound of /x/ in the word “to hate” /xɤn/ and the clear vowel diphthong of the word “to love” /ai/. The /x/ sound may be associated with the snarling sounds of animals and other rough sounds made due to angry emotions and thus be associated with this sound. If this is true, then it means that if this experiment were to be repeated, that the categories would need to be further divided into more varied groups.

3. the abysmal accuracy rate of 40% of the pair “rough/smooth” (粗/滑)as opposed to the average accuracy concrete word pair rate of 64.4% may be due to the word “rough” being pronounced / tsʰu/ and thus sounding similar to the English word “smooth” in that both words have an sibilant sound near the beginning. 
4. Finally, the pair “dog/cat” (狗/猫) was taken by some subjects to be a trick question; many guessed correctly at first, attaching the word “cat” to the sound / mɑʊ/ but immediately withdrawing the guess and opting for the other combination, which was wrong. Despite my telling subjects to use their “gut feeling” they still experienced this reverse psychology and this may thus have affected the results.

5. Somewhat relating to point one above, I realised in retrospect that the languages of Japanese, Korean and Vietnamese have large amounts of Chinese vocabulary and that subjects may have had knowledge of one or more of these languages and thus given them an unfair advantage, skewing my results. 

6. Although this study did not specifically look into the tones of the words, it is quite certain but not proven that they impacted the subjects’ perceptions of the word pairs and are a part of sound symbolism. Further study may be needed in this area.

7. The words I chose had not been random, these were words chosen because of my believing them to be either with or without sound symbolism and if given more time, I would continue this study with more vocabulary and more subjects.

Conclusion:

In conclusion, based on this investigation, Mandarin Chinese does have sound symbolism and that there is slightly more sound symbolism in concrete than abstract words. This proves my hypothesis, though not to the level that I had envisioned it. I have discovered that sound symbolism exists in Mandarin and people, myself included, might not even realise its influence upon our intuitive feelings about sound and meaning. Despite these findings, questions remain and questions have sprung up. Will this degree of sound symbolism hold true if all nouns, verbs and adjectives were to be tested? Does emotion allow abstract qualities to have sound symbolic values attached to them? Do certain older varieties of Chinese such as Cantonese and Hakka have more sound symbolism than their younger cousin Mandarin? For now, within the context of the survey questions and the surveyed subjects, it can be assumed that there is a slight relation between sound and meaning in Mandarin Chinese.

Bibliography:

Köhler, W. (1929) Gestalt Psychology. New York: Liveright.

Magnus, Margaret. (1997-2001) Margo’s magical letter page.
            Retrieved August 13th, 2007 from http://www.conknet.com/~mmagnus/ 

Mathews, P.H. (1997, 2005). Oxford Concise dictionary of Linguistics. 

Oxford University Press: England

Pinker, Steven. (1994). Grammar Puss: the Fallacies of the Language Mavens. 

Retrieved August 13th 2007 from http://camba.ucsd.edu/~bakovic/ll/grammar_puss.html

Ramachandran, V.S. & Hubbard, E.M. (2001b).
            Synaesthesia: A window into perception, thought and language.
            Journal of Consciousness Studies, 8(12), 3 - 34.

Sabaneev, L. & Pring, S.W.(1929). The Relation between Sound and Colour.
            Music & Letters Oxford University Press; England