Capaldo, Alexander Schultz

An Examination of the Sound Symbolism of Vowel Characteristics in Brand Names, Briana Caesara

Let the Writer Write and the Reader Read: The Infulence of Ourlipo on Computer-Generated Poetry, Erien Greenhalgh

Capaldo, Alexander Schultz

(Back to top)

PERSONS OF THE DIALOGUE:

HURST, ISLAM, SCHULTZ, SEELENFREUND, AND CAPALDO

The prologue takes place in the outer commons of Willville, and afterwards in a restaurant near Bear Creek apartments, in the year 2011..Hurst and Islam are mentioned on Schultz’sFacebookas having come from Denver to room with Seelenfreund and himself. Little is known about them beyond the fact that Hurst adhered to a strong version of CU’s focus, as a walk-on candidate for the football team. He is thus a likely person to be interested in whether his perception that he will make the team would confirm that it is true that he will make it.

HURST. ISLAM.

Hu. Are you back from work for long Islam, or do you have to return to handle the equipment?

Is. I’m done. No more practice today. Coach gave the team the afternoon off. I was actually looking for you back at the room. Care to go to McDonald’s?

Hu. After all the cookies you have eaten? Well, I can hardly say no. I have most of my day free now.

Is. I had a crazy time with the team today. A couple guys got kicked off after they ditched a drug test, which counts the same as a failed one. CU now has a converted wide receiver at corner!

Hu. That is just sad. Those boys need to get it together.

Is. It did not make clean up any easier, that is for sure. When are you going to try out? Embree could probably use you.

Hu. Well, it is a question of my knee right now. I don’t want to mess it up any more. I wish I just knew whether or not I should get surgery.

Is. Well, if you heard Schultz and Capaldo going at it before, you might think you could answer that question pretty easily.

Hu. How so?

Is. Well, when we were at McDonald’s earlier, they started bickering. Capaldo said something about knowledge being determined by particular people, and Schultz erupted.

Hu. Capaldo? Really? Last I heard of him, he was getting into fights over at the Center for Community. Is the idea that if I come to think that I can play without surgery, then that would be true?

Is. Sounds nice, I must say. I thought that was Capaldo’s point. But Schultz did not seem to accept it, and they parted rather angrily.

Hu. I have plenty of time on my hands. Can you tell me about what happened?

Is. I would, but my mind is frazzled from today’s practice. I don’t want to misrepresent either party in the discussion. However, I managed to videotape the diatribe, and I put it on YouTube. It’s starting to get some views, want to watch it?

Hu. Sure, pull it up on your iPhone while I get a McRib. Care for one?

Is. I’ll have a McDouble. And some fries.

The scene of the dialogue itself is a low-class restaurant in Boulder and the date shortly before the 45-2 drubbing CU suffered against Oregon on October 22, 2011A.D.Seelenfreund is a distinguished mathematics and M.C.D.B. major from Denver. Capaldo, an ex-high-school football player, is a friend of Islam, who is present in the background, talking with some friends. Capaldo is of a lower class known as Freshmen, and is at first eager to impress.

SCHULTZ. SEELENFREUND. CAPALDO.

Sch. Tutoring today was fun. Remember that cheerleader I told you about—the one I think is cute, even though she is in a sorority? She is doing a paper on relativism. She asked me what I thought about it!

Se. Really? Did she think relativism was true or something?

Sch. She thought so for a second, and I didn’t want to give it away immediately. So I told her why I think people fall into such patterns of thought—namely, because it is hard to tell unintelligent people that they are just wrong about something. So relativists try to play knowledge off in a conciliatory manner, saying that truth is what people think it is.

Se. But as a philosophy major, you seem to have no problem telling people they are wrong.

Sch. Well, I just have an advantage with regards to epistemology, so to speak. It is my duty! Plus, I love telling relativists that they are wrong, because they always have to agree with me that I am right.

Se. Funny, that. I am sure society thanks you.

Sch. Joking aside, I did manage to get her to go along with relativism for a minute. Want to hear the first argument I gave her?

Se. Of course I would. We have plenty of time, seeing as Islam takes forever to eat.

Sch. I said something about different people having different experiences, and given these unique experiences, it is impossible to make claims about what their experience really is. Given this subjectivity, it is just as impossible to criticize their beliefs about what is and is not true—because, after all, their beliefs arise from their unique experiences.

Se. Hilarious! And she bought it?

Sch. Well, I am her tutor. It is hard for me to recognize, at times, when instructors mislead me during their lectures—after all, they are the ones imparting new information to me, especially when I don’t do my readings. So I do not fault her much for finding what I said to be true.

Se. Still, relativism always struck me as outlandish. I mean, who can think that there are no objective facts? I must say, my field would never allow that. Imagine if, as a doctor, I could not prescribe medicine before I got my patient to believe it would work! Absurd.

Sch. Well, once I argued for the other side, she quickly changed her viewpoint.

Se. What did you say?

Sch. “Of course, if you start believing that truth is just what people think is true, you would not be a cheerleader if you stopped thinking you were one.” She freaked out then, muttering something about how she had worked too hard to make the squad to let it be taken away so easily.

Se. That was well refuted, indeed.

As the two were speaking, Capaldo had tried many times to voice his own thoughts on the matter, but he had been waved off by the two Sophomores as unlearned in the terminology. When they paused after what had last been said, however, his eagerness to please had been replaced by anger becoming of a coiled, wild beast, and remnants of Mcnuggets seemed to fly from his mouth as he erupted.

Ca.Are you insane? What nonsense you two have paraded around as argument. If you really want to convince someone that relativism is false, don’t just agree with one another until you reach what you apparently have already concluded. You know very well it is easier to agree than to actually oppose one another until you reach a valuable end. Why don’t you give a real refutation, and this time not to a friend who will simply accept whatever it is you seek to determine before the argument has even been tested fairly, but to me, for I won’t accept such nonsense from you.

Schultz was visibly startled by the outburst, and when he did reply, his words were somewhat unsure.

Sch. Don’t be so hard on us, Capaldo, for if Seelenfreund and I seemed sloppy in our refutation, it was not because we did not take the subject seriously. In fact, we have discussed this before, and so we were simply reaffirming what we both determined earlier; namely, that the idea that knowledge depends on what people think about it is false.

Ca. I will not fall into your trap. Even now, I see how you seek to convince me of this with friendly words, but your argument now is just as fraught with difficulties as the one before it. You tell me that you know from some earlier discussion that such an account of knowledge is false, and yet again you give no real reasons towards this view. It seems as though I must come to your rescue, for I know the real reason behind your lack of ability. It is because what you seek to say, that truth is not determined by the individual, is the weaker view, whereas the strength of the matter really lies in what you term ‘relativism.’ For it is really the case that what people think is true is true for them, and what you think is true is true for you. There is no objective element of truth. You call yourself a philosophy major; it sounds to me like you should read up on Protagoras.

Sch. Well Capaldo, if you insist upon the faultiness of the conversation I just had with Seelenfreund, I feel as though I must turn to address what you now tell me. I find it fitting to be asked to come to the defense of the refutation of relativism I have already put forth, and will gladly examine with you my reasons for holding it. In the process, we will see if your own viewpoint may yet be changed; I certainly see many objections that will interfere with your desire to maintain that relativism is correct.

Ca. Yes, we will see about that—I think you will find that I am well versed in the ideas of Protagoras, and you will soon be forced to hold that all you thought about how truth operated does not apply to any beyond yourself.

Sch. Come then, tell me this. What is it specifically that Protagoras tells us about knowledge?

Ca. That knowledge is perception.

Sch. Ah, an interesting doctrine. I will personally draw on my understanding of Plato in this matter, as I believe he is well versed in arguments against these types of theories.

Ca. Very well. What does he, or you, have against the theory put forward by Protagoras and myself?

Sch. Well, my first problem is that it seems to require that no person is wiser than another in any matter. It seems intuitive to say that a football player is wiser when it comes to the rules of the game than somebody here who flips burgers. Yet if we are to hold, as you say, that knowledge is perception, then no man could judge another’s experience better than they could their own. In such a situation, the only person able to determine the knowledge of a particular matter, and here we have football rules, would be each person for their own case. Generalizing on this principle, it seems that in all matters, the judgments of all persons are correct—and so there is not one who is wiser than any other. We are all just correct. Is this really what you are seeking to hold?

Ca. Why, you have supplied me with yet more ammunition to back up my theory, and it seems as though you have done so unwittingly. Yes, Schultz, this is what I am seeking to hold; I want to say that what people think to be true is indeed true for them. You say that a football player is wiser than a burger-flipper when it comes to the rules of football. I say not wiser in the sense that you mean, but better. These two come from vastly different experiences, experiences that are unique from all others. All the burger-flipper knows of the rules and understands to be true is all that will affect his experience of his life at that moment. Thus, his conception of the rules is true for him, because that truth holds throughout his own life and his own experiences. Nobody else can know what his own experiences are; they can only come to affect them. They cannot change the fact that it is his life that creates his own truth, but they can change the truths that are in his life. The fact that the truths change matters not; for if he comes to a new, comprehensive understanding of the rules, then that is the truth of his world at this new moment, but it does not change the truth of his past. For his truth is whatever is affecting his experience at that time, in his own world-view, and so it is impossible for any to say that his perceptions are wrong for him, just as much as it is impossible for us to say that we know what it is like to be him.

As for your wisdom objection, I would say that there is still a sense in which the football player is still wiser, simply because wisdom is not as closed a word as you may have thought it to be. You might think wisdom relates to more accurate and correct knowledge about a subject. However, I would say that wisdom really means the ability to change appearances to a state that isbetter, which would lead a man to think things that are good. I think it is uncontroversial to say that some things are better to think than others, but that is not to say, as you seem to think, that they aretruer. Take our football player, for he is wiser in the sense that he will be better able to follow the rules, and more likely to foster agreement in his explanation of them. But I hold that our burger-flipper is in no way compromised in histruthwhen he thinks a touchdown is worth five points, because in his mind there is nothing jumping out at him and saying ‘WRONG’ when he holds this view; in the same sense it is impossible for someone to sincerely say “I believe P, but P is false.” As such, his perception here cohabitates with his current experience, and it is still true for him that a touchdown is worth five. However, this view might be bad in the sense that there are others that might threaten him about his truth, lambasting him as stupid or unversed in the importance of football, and so it is still within the power of a relativist to call his truth bad, and the football player wiser on the matter.

Sch. It seems that the only way for something to be false under relativism would be if someone were able to have a perception and, in that moment, use it to falsify that very same perception they are having.

Ca. Exactly. But that is impossible, which makes perfect sense, given that truth depends on the individual’s experience.

Sch. So could one frame your argument as one with premises and a conclusion? Tell me if this gets to what you are saying, namely, that a) nobody knows what another’s experience is like, b) we use our experience to form perceptions, which are necessary and sufficient for knowledge, c) knowledge is sufficient for truth, and finally d) thus, truth depends only on the individual, because others cannot say what a truth may be because we do not know what other individuals’ experience is like.

Ca. That is a good view of part of the argument, to be sure.

Sch. Well, if I were to give you an example of a perception that would, in the same instance of having it, serve to falsify that perception, would you then retract your theory of relativism?

Ca. Honestly, I do not see why I would have to.

Sch. Let me explain myself then. Would you say that you could know something that is false?

Ca. In the sense that I would know that a particular thing is false, or in the sense that there is something I know, thing X, and thing X is false?

Sch. The latter.

Ca. I would say no, considering the views of Protagoras and myself. For if I have come to perceive it, and thereby know it, then it must be true for me.

Sch. Good. So, in such a sense, could you have a false perception?

Ca. It would seem as though I could not.

Sch. And by the same line of thinking, could you have a perception that, by being true, makes itself false?

Ca. Well, if my perception were to make itself false, then it would again be false for me, and so again I say that such a perception is impossible.

Sch. So it seems as though relativism requires you to say that there are no such perceptions.

Ca. And indeed I would hold that there’tany.

Sch. Let us see about that then. Would you say that there are some people, like myself, that think you and Protagoras are wrong?

Ca. Evidently. Furthermore, I would say that such a belief is true for you, although I wish you would come to see my side, for it is certainly the better one. It allows us to engage in a search, not for the truth, but for the best possible state of affairs—one that eliminates the judgments of others, but allows for the possibility of the betterment of their viewpoints.

Sch. That is an interesting idea, Capaldo, but right now I am concerned more with whether or not it is right. I would gladly search for this best possible state with you, but first I must come to believe for myself that such states can be found without appeal to objective truth. So it seems as though you must continue to convince me, through refutation of the arguments I present to you, if we are to come to that.

Ca. Very well. Please continue.

Sch. So relativism is wrong, by my standards.

Ca. That’s what you think.

Sch.Well put. Now, you believe in relativism, do you not?

Ca. Yes, I think that relativism is right.

Sch. So, in your head, you have a perception of both relativism being right, and that I think it is wrong, correct?

Ca. I fail to see how twisting language into complicated locutions is going to help you, but yes.

Sch. Just wait. Doesn’t your perception then require that a) I am correct in my perception that relativism is wrong, b) relativism is right by your perception, and so I am correct in thinking that, for me, relativism is wrong and thereby c) I am correct because relativism is wrong and I am correct because relativism is right? Is this not a demonstration of a perception that makes itself false by being true?

Ca. Ah, I see what you are trying to do yet again! Through trickery and wordplay, you seek to turn my theory against me. But you have again failed your own directive, because your reasons have failed you for yet another time. You were right in saying that if there were a perception that, by my having it, were to make itself false, then relativism would not stand. But you have not done this—you have just shown me a perception that leads itself to contradictions. And there is nothing wrong with that.

Sch. Do you not know, though, that everything follows from a contradiction?

Ca. That may well be true for you, but certainly not for me. You have given me a contradiction, but I see no purple dinosaurs stealing my Big Mac. The laws of physics have not broken and no French Fries float tantalizingly by, suspended in midair. I think your demonstration has just shown, if anything, that your viewpoint of relativism being false is worse than my viewpoint of it being true, because yours falls into contradictions but mine does not.

Sch. You are ridiculous. Have you not heard of the law of noncontradiction?

Ca. There you go again, trying to usead hominemstrategies to vouch for your theory. No, I have not, but I doubt it will save you either.

Sch. It simply states that contradictions cannot exist in the world. Do you believe something can have both property A and not A?

Ca. No, I do not.

Sch. Well, no pun intended, but you just contradicted yourself. Earlier, you said the perception was a contradiction, but you saw no problem with its existence. Yet now you say that nothing can be both A and not A.

Ca. Again, you are trying to surgically manipulate my words with a wrecking ball, and it does not lead to success in building a refutation. I do not believe that something can be A or not A, but that is my truth. The contradiction you gave me was your truth. For me, relativism is right, and by it I would say that your perception is restricted to yourself. So your claim that I believe in a contradiction is not true for me.

Sch. Perhaps I must try another track to get to you.

Ca. How many tracks must you try until you accept that my position is better? I am growing weary of this, and my body yearns for harsh physical contact in which a small ball is involved.

Sch. Allow me one further attempt, and if your reply convinces me that it does not hold, then I will come to your view. However, you must accept that relativism is wrong if I can show you that my refutation does survive. If neither of us finds agreement, then we will part and end this discussion.

Ca. Well, what is it? What fallacy of logic do you seek to employ now?

Sch.Merely this. I am going to accept that relativism is true.

Ca. A fine decision!

Sch. Now let us look at whether claims we can make underneath this theory are compatible with it. I can certainly say that McDonald’s French fries smell good, can I not?

Ca. Of course, with the implicit understanding that such a claim is only true for you.

Sch. So now, can I claim that it is true for me that McDonald’s French fries smell good?

Ca. Certainly, again there is no problem. I am glad that you have accepted relativism, as you said.

Sch. It is an assumption. Can I say that it is not true for you that it is true for me that McDonald’s French fries smell good, if that it is what you perceive?

Ca. Your wordplay has about gotten me to the point of spitting my Dr. Pepper in your face.

Sch. Your mouth already forced the future doctor out of the conversation. My last problem with relativism is that it restricts us to our own worlds of truth. My world cannot influence yours in manners of truth, but merely “make it better.” However, if we are going to look at relativism through such an absolute lens, then these judgments that are supposed to be true for only us cannot be relativized into others’ worlds of truth. This is going to cause a problem. It seems that it is not relative that it is true for me that McDonald’s French Fries smell good. No matter who thinks such a thing, no perception can make it false that it is true for me that the Fries smell good, because such a perception, while it may be held by them, does not depend on them—it depends only on me. But your theory also wants to say that even that kind of knowledge is relative, because all knowledge is perception, and so relative to the perceiver. But these claims are inconsistent with each other, and there you have your contradiction that does not depend on “truth for.” Now, either tell me that things have properties they do not have, or tell me relativism is wrong.

Ca.I’ll tell you that your method, Schultz, is one of theoretical wordplay that does not get to the heart of what I am saying. You seem to just not get it. And like we agreed to earlier, since you have not convinced me, I am going to leave you to rot within your bad viewpoint. I tried my best to convince you to switch to the better one, but it seems not everyone is willing to see the light.

Sch. You say you are not convinced by me, yet you will not give me one reason as to why?

Ca. No. You do not deserve it. Ask your doctor friend, I’m sure he will continue to agree with you.

Capaldo storms out of the restaurant, not even bothering to bring his McDonald’s Monopoly® game pieces with him. Schultz and Seelenfreund turn to each other.

Se. That guy was outlandish.

Sch. Honestly, I cannot wait to run into him again. My arguments were more than convincing. Next time I’m taking notes.

Se. I think Islam has it on video.

Sch. Really? I didn’t think he could eat and pay attention to other people at the same time.

An Examination of the Sound Symbolism of Vowel Characteristics in Brand Names, Briana Caesara

(Back to top)

Introduction:

Brand names are everywhere in our society and there are more brands trademarked every year. Most people are aware of the link between successful products and catchy brand names, but are much less clear about what makes a brand name effective. After all, who would have thought that Kleenex would be so successful as to become a generic word for tissues or that things called iPods would completely change the way people listen to music. However, while that kind of success also requires a quality product, an ineffective name can doom a product from the start. This is why international translations of product names are such tricky business, because simply keeping the current name often leads to unintended connotations. Some of the best examples of this are products marketed in Germany that contain the word “Mist” in the name, because this translates roughly as “dung.” However, while most international branding issues are a matter of unfortunate translations, an awkward or inappropriate sounding name can have a similar effect as well, and this is where phonetics enters the picture.

This paper will look at the perceptual effects of vowel characteristics in brand names, to test how differences in vowels can affect the perceived attributes of the product. The focus is on replicating and expanding past linguistic and market research into sound symbolism, which found that having different phonemes in a word can alter people’s perceptions of a word’s referent. Of particular interest for this study is the article by Richard Klink:Creating brand Names with Meaning: the Use of Sound Symbolism (2000).

Background

1.1: Sound Symbolism

The Oxford Dictionary defines sound symbolism as “the partial representation of the sense of the word by its sound.” In linguistic study, this is the idea that individual phonemes carry meaning in and of themselves. This very general definition includes all degrees of this linkage, from onomatopoeia to the conventional association of meanings with phonemes such as the 'gl' of “glimmer,” “gleam” and “glisten.” The focus of this paper falls between the two extremes, centering on the idea that different vowels and consonants can consistently imply physical or visual properties about an object separate from the semantic meaning of a word.

One of the earliest English-language studies in this area was performed by Robert Sapir in 1929. Sapir (1929) tested the “symbolic magnitude” of differences in vowel and consonant characteristics, regardless of their presence in meaningful words. To do this he studied 464 American students and found that in 80% of the cases they ascribed large size to nonsense words containing the low back vowel /ɑ/ and small size to those containing the high front vowel /i/. Given these findings, Sapir (1929) proposed that this was due either to differences in the frequency of various vowels or the kinesthetic association of the “spatially extended gesture” (the more open position of the tongue and mouth for back vowels) with larger volumes. While this experiment contrasts both frontness and height, further studies focused on and found support for a front/back vowel distinction for a number of other attributes, such as fast-slow, weak-strong and thin-thick (Klink, 2000). However, size remains one of the most recognized forms of sound symbolism.

One of the most famous of these sound-attribute relationships was that found by the German-American psychologist Wolfgang Kohler, who did work on sound symbolism relating to shapes. Kohler (1929) found that people would reliably match the nonsense words “Takete” and “Maluma” to images of jagged and curved shapes, respectively (cited in Nielsen & Rendall, 2011). In additional studies, he found similar effects with the words “Kiki” and “Bouba”, with the former attributed to pointy shapes and the latter applied to rounded ones (Kohler, 1947, cited in Nielsen & Rendall, 2011). Kohler's original stimuli imply that the Takete-Maluma phenomenon is due to the differences in the consonants rather than the vowels, since the first experiment contrasted stops and sonorants rather than the vowels. However, the Kiki-Bouba experiment contrasted both different types of stops and front/back vowels, so the effect could be due to either. Whichever segments are responsible, these experiments clearly show that sound symbolism can alter people's perceptions of attributes other than size, which has interesting implications for real world applications of this effect, including in creating brand names.

There has been limited research into the applications of sound symbolism on marketing and creating brand names. However, because many brand names are simply invented nonsense words, the attributes associated with their phonemes should have more effect on people's perceptions than those that must also contend with the linguistic meaning of a word. Thus, project developers and marketers would benefit highly from knowledge of sound symbolism when attempting to design a suitable name for their product. Relaying product information through sound symbolism within the brand name is the focus of the 2000 study by Richard R. Klink.

1.2: Creating Brand Names with Meaning: the Use of Sound Symbolism by Richard R. Klink.

Unlike earlier marketing research, Klink (2000) looks directly at the question of whether the phonemes of a brand name can influence people's perceptions about different products. The study compared front and back vowels, stops and fricatives, and voiced and voiceless stops and fricatives within imaginary brand names to test whether their phonetic characteristics have a consistent effect on people's assumptions about the product attributes. Klink (2000) examined thirteen attributes for vowels: size, darkness, mildness, thickness, softness, speed, temperature, bitterness, masculinity/femininity, friendliness, strength, weight and attractiveness. He also looked at speed, size, darkness, sharpness, softness and masculinity/femininity for consonants, but that part of the study is unrelated to the purpose of this paper. The hypotheses for the vowel comparison in Klink (2000) was that the products with brand names containing front vowel sounds are perceived as smaller, less dark, milder, thinner, softer, faster, colder, more bitter, more feminine, weaker, less heavy and prettier than those with brand names containing back vowel sounds.

The procedure in Klink (2000) was visually-based, with 265 American undergraduate students given a booklet that informed them that they would be evaluating potential brand names. Within the booklet were questions relating to each of the hypotheses, such as “Which brand of Ketchup seems thicker? Nidax or Nodax” (p. 11). Each attribute in Klink (2000) was tested with four different word pairs that contrasted the letters “i” and “u,” “i” and “o,” “e” and “u,” and “e” and “o” respectively. The initial word pairs consisted of both one and two syllable words, however during pre-testing the majority of one syllable word pairs were eliminated due to their resemblance to real words or existing brands. Within the questions for Klink (2000), the order in which imaginary brand names appeared was also altered and the order of the questions was randomized.

Klink (2000) found that all hypotheses for the perceptual effects of vowel frontness on product information were supported. Thus products with brand names containing front vowels were perceived as softer, faster, thinner, smaller, less dark, milder, colder, bitterer, more feminine, friendlier, weaker, less heavy and prettier than those with back vowels in their names. However, while this has numerous implications for product marketing, from a phonetic standpoint there is a significant issue with the study's procedure: the use of written brand names rather than oral examples. While this was a deliberate choice by Klink (2000) in order to make the study more efficient and more closely replicate people's interaction with brand names in the real world, it also undermines the veracity of the study's phonetics. As any phonetician knows, the relationship between English orthography and pronunciation is extremely limited, so the reliance on written brand names leaves the actual vowel sounds in the stimuli open to interpretation. While the consistent results found by the study imply that the subjects were coming up with similar pronunciations for these word pairs, it is difficult to know whether the results are actually due to the intended comparison or another unknown factor. Given the wide variety of sounds in English made by the letters “i,e,o, and u,” the perceptual differences could have been due to a number of factors, including rounding, height and frontness. It is these phonetic uncertainties that this paper attempts to address.

Theory and Hypothesis:

The purpose of this paper is to begin looking at some of the unanswered questions raised byCreating Brand Names with Meaning: the Use of Sound Symbolismby Richard R. Klink (2000). First, using a similar procedure but giving the subjects recorded brand names rather than written ones should eliminate the variation in brand name pronunciation. Second, by using word sets that contrast vowels from the International Phonetic Alphabet rather than orthography, the phonetic characteristics of the vowels are kept more precise. Finally, the use of front-back vowel comparisons and high-low vowel comparisons will help to determine the actual cause of any sound symbolism that appears.

2.1: Front-back Vowels

The hypotheses for this portion of the study are that products with brand names containing front vowels will be perceived as smaller, thinner, faster, colder, friendlier, weaker and prettier than those with brand names containing back vowels. The reasoning for this is based on the earlier linguistic research that supports this distinction, as well as the results of Klink (2000). Thus, this paper expects:

H1 (a-g): Products with brand names that contain front vowels in comparison to back vowels are perceived as (a) faster, (b) smaller, (c) colder, (d) thinner, (e) weaker, (f) prettier, and (g) friendlier.

2.2: High-Low Vowels

The hypotheses for this portion of the study are that products with brand names containing high vowels in comparison to low vowels, will not be consistently perceived as any smaller, thinner, faster, colder, friendlier, weaker, or prettier, nor the reverse. The reasoning for these hypotheses is based on the fact that the two possible theories Sapir (1929) offered for the size symbolism of front-back vowels, conflict with each other when applied to the phonetic characteristics of the high-low vowel distinction.

Sapir (1929)'s first possible explanation was based on the relative frequencies of front and back vowels. The major difference in frequency between front and back vowels is the relative height of the second formant, with [i] having the highest value. Ohala (1997) argues that the size symbolism attributed to this difference has biological roots, based on Morton (1997) who studied the sounds made in competition between bird and mammal species. This study found that the aggressor made low frequency sounds, while the submissive animal emitted high pitched sounds (as cited in Ohala, 1997, p. 2-3). Morton postulates that this difference is due to the fact that the absolute frequency of a sound if inversely related to the length of the vocal tract (as cited in Ohala, 1997, p. 3) Thus, larger creatures emit lower frequency sounds and smaller creatures emit high frequency sounds. Ohala (1997) argues that the cross-language appearance of the front-back vowel size symbolism for humans is thus due to a biological hold-over from our more animalistic days.

In contrast, the major difference between high and low vowels is the height of the first formant, with low vowels having the higher frequency. Thus, if frequency is the primary factor low vowels should find themselves linked to the same attributes as front vowels. The situation is complicated by the fact that while the back vowels /u/ and /ɑ/ theoretically have the same second formant height this is not always the case. This may particularly be true in this study, as the word pairs were recorded by someone with a California accent, and thus a generally fronted /u/. Also, even on the IPA chart /i/ is more fronted than /æ/ so the difference in the second formants of high and low vowels could affect perception. Thus, while the theory suggests that front and low vowels should have similar effects on the perceived attribute, the case is not cut and dry.

To further complicate the issue, Sapir (1929) also theorizes that the front-back difference was that the size symbolism was due to the difference in the mouth and tongue position used to create the vowels. Thus, front vowels are perceived as smaller because the vocal apparatus is more constricted when they are made in comparison to back vowels. If this is the primary factor than front vowels should have comparable effects to high rather than low vowels, because air flow is more constricted when making high or front vowels. Thus, these two theories contradict each other when applied to the low-high comparison, so there should be no pattern, particularly given all the other complications mentioned earlier. If a pattern does emerge this will have interesting implications for which of the theories are correct. Therefore, this paper expects:

H2 (a-g): Products with brand names containing high vowel sounds as opposed to low vowels will have no noticeable pattern of perception as relating to (a) speed, (b) size, (c) temperature, (d) thickness, (e) strength, (f) attractiveness, or (g) friendliness.

The Study

3.1: Procedure

The subjects for this study were ten native English speakers. Each subject was emailed a link to a Qualtrics Survey and asked to help evaluate potential brand names by completing the questions. The survey consisted of seven questions relating to the seven tested attributes. The questions were in the style of “Which brand of X sounds X? A, B, C or D” with A, B, C, and D being made-up brand names that contrasted /i/, /u/, /æ/ and /ɑ/, and X being the superlative of the attribute being tested, such as “smallest.” In the survey, the order of questions was randomized for each subject and the order of the contrasting vowels was different for each question.

3.2: Stimuli

The survey required 28 words, consisting of seven sets each of four contrasting words. These imaginary brand names were created in the pattern of V̥-stop_ [{V̥-stop} coda], with the _ contrasting /i/, /æ/, /u/, and /ɑ/ for each set. Keeping all other phonetic characteristics in the words uniform was done to limit the effects of allophonic rules on the experiment, such as the lengthening of vowels before a voiced sound, so that only the hypothesized characteristics were contrasted. In keeping with this goal, the second syllable of the word always ended with a liquid, nasal, or voiced stop so as to reduce sound variety. The greater variety in the final consonant was due to the need to avoid resemblance to actual words, while keeping the characteristics of the consonants adjacent to the vowels as similar as possible. For a complete list of brand types and their respective attributes and the words sets for each question, see Appendix I on page 19.

3.3: Results

For each question, the responses were aggregated to their overall category, such as front, back, high or low vowels, and then compared to reveal the larger pattern. To see a chart of all results and the individual breakdown of responses, see Appendix 2.

Front vs. Back Vowels:As shown in the graph above, the majority of the H1 hypotheses were supported, although the level of the correlation varies greatly by attribute. There were also 3 exceptions, with attractiveness, friendliness and temperature showing no clear connection between vowel frontness and the attribute. Thus H1 (a), (b), (d), and (e) were supported while (c), (f), and (g) were not.

High vs. Low Vowels:As shown in the table, the majority of the H2 hypotheses were not supported, as a correlation was found between height and attribute, including some of the attributes where the front-back distinction was not supported. There were only 2 exceptions to this, with strength and prettiness showing no clear connection, although the correlation for temperature is slim as well. Thus H2 (f) and (e) were supported, while H2 (a), (b), (d), and (g) were not, and (c) was only slightly not supported.

Discussion

While there are connections found in the aggregated results, when individual vowel pairs are compared within the responses, the results become more complicated. In some cases, they show the same connections seen in the aggregated results and in others the individual correlations appear to be contradictory. Yet, for all attributes examining the results for the individual questions allows one to begin theorizing about the elements that caused these correlations for each attribute, and illustrate the areas in which more research is needed.

4.1: Speed

Speed has one of the stronger correlations between the attribute and vowel height and frontness, with both high and front sounds being perceived as faster. Looking at the results for each individual vowel, one can see that the correlation appears to move in a continuum, from /i/ as the fastest sound to /ɑ/ as the slowest, with the high back and low front vowels being chosen the same amount. The airflow constriction appears to be more important than the formants, as low vowel and back vowels have similar response percentages, but inverse formant heights for F1 and F2 (front vowels have high F2, while high vowels have low F1). However, in this case the first formant may be the driving factor, as a fronted /u/ has a similar F2 as /æ/, both of which are lower than the F2 of /i/. To discover whether this pattern is a matter of accent or actually related to the difference in constriction, further research using different stimuli is required.

4.2: Size

Size also has a strong correlation between vowel position and the perceived attribute. In this case, while high and front vowels are strongly correlated in the aggregate results, the individual results show that it is the high front vowel that is most strongly connected with smallness. The single choices of /u/ and /æ/ imply that height and frontness are important separately as well as together for perceived speed, but without additional data one cannot know if those responses were simply aberration. Like the previous data that implies that formants are less important for this attribute because of the inverse nature of F1 and F2 for front and high vowels, but the strong emphasis on /i/ combined with possibly /u/ fronting may mean that high F2 and thus high pitch is actually the most important factor. This would align with the theory that the perception of size is due to the fact that smaller creatures make higher pitched noises.

4.3: Temperature

Temperature is unique amongst the attributes in that one respondent felt that all words sounded equally warm and so declined to answer. There is also very little correlation between the vowel sounds and the perceived attribute, although high vowels appear to sound somewhat warmer than low vowels which may imply that constriction is the deciding factor. However, the most common responses for this question were the high back /u/ and the low front /æ/. These vowels are polar opposites in both constriction and formant frequencies, so the cause of this pattern is unclear. One possible explanation is that the relationship of the vowels formants to each other matters, as the two vowels have low-low and high-high F1s and F2s respectively.

4.4: Thickness

Thickness shows the strongest correlation of any attribute, with 90% of people choosing the high back /u/. This shows a strong correlation of both height and frontness to perceived thickness, with an emphasis back vowels sounding thicker. In this case the formant frequencies may be the most important factor as back and high vowels both have lower formants than their counterparts (F2 and F1 respectively), but inverse amounts of constriction. However, without more data it is difficult to say whether the pattern holds true when comparing only the low or front vowels, or whether it is the combination of those positions that is important.

4.5: Strength

Strength is the first attribute that shows a correlation for frontness but not height, with both back vowels strongly correlated with strength. This implies that it may be a low F2 that most affects perception of products as stronger because the back vowels agree on this trait but not on constriction or F1. A low F2 representing strength would be in line with the theory stated by Ohala (1997), which says humans perceive lower frequency sounds as belonging to larger, and thus stronger, animals.

4.6: Attractiveness

Attractiveness has the weakest correlation between sound and attribute, with no pattern in the front-back distinction and low vowels only slightly perceived as prettier than high vowels. While the latter could show a slight correlation between F2 and perceived prettiness, without more data one cannot say whether this is simply a false pattern. If no correlation exists, this may be due to the subjectivity of the tested attribute, as people have different opinions of what is attractive. However, Klink (2000) found a correlation between front vowels and prettiness, implying that subjectivity is not the answer. This may be due to the difference in procedure, as people could have been affected by the attractiveness of the written name, but it worth further investigation. Further research into whether a pattern occurs when vowel characteristics are tested individually in front-back and high-low recorded word pairs or with a greater number of subjects would be the place to start, as those are two of the major differences between Klink (2000) and this experiment.

4.7: Friendliness

Friendliness also shows no correlation between front vowels and perceived friendliness, but there is a strong connection between low vowels and the attribute. These results imply that a high F1 is the most important factor, as /æ/ and /ɑ/ differ in constriction but both have high F1s. While this is a clear pattern, this is the opposite of what was expected. Klink (2000) found a front-back correlation for this attribute with the high front vowel /i/ as the friendliest sound, possibly due to its similarity to smiling. In contrast the highest response for this experiment was the low back /ɑ/, which requires the opposite mouth shape. This difference may be due to the shift in procedure from written brand names to recording, but is worth further investigation.

Conclusion

While the strength of the correlation varies depending on attribute, the results of this study show that the sounds of brand names do have a significant effect on how their products are perceived. This is particularly important for those products where speed, size and thickness are important traits because they had the strongest correlations and so the name would create stronger perceptual effects. However, every attribute except attractiveness showed a pattern of some kind, and even that did have favored and disfavored vowel sounds. Thus, vowel sounds are not something that marketers can afford to ignore when creating new brands.

Interestingly, the results of this experiment do not match those of Klink (2000) in every case. This may be due to differences in procedure, particularly in the use of discrete sounds rather than orthography to make the comparisons. It may also be due to the accent of the person recording the sounds, as a different dialect could theoretically yield different results. While a true comparison would require more data, these results imply that there can be a significant difference in the perception of words based on whether they are heard or seen, although in both cases the vowels do have a perceptual effect.

Further research is necessary to determine whether the patterns observed in this study hold true across a greater number of responses. It is also required to determine whether the unexpected results were flukes, caused by the dialect, or due to procedural differences. Either way, the results of this experiment show the need for future studies into how vowel characteristics influence perception and how to explain the patterns that are found. Research into the effects of consonants would be beneficial as well, because this study examined vowels only in conjunction with voiceless stops, and differences in place or manner of articulation may effect perception as well.

References

"sound symbolism". (April 2010) . In Oxford Dictionaries online. Retrieved from .

Klink, R. (2000). Creating brand names with meaning: the use of sound symbolism.Marketing Letters,11(1), 5-20. Retrieved from Business Source Complete Database via Chinook.

Nielsen, A., & Rendall, D. (2011). The sound of round: Evaluating the sound-symbolic role of consonants in the classic takete-maluma phenomenon.Canadian Journal of Experimental Psychology,65(2), 116. doi: 10.1037/a0022268 Retrieved from Academic Source Premier via Chinook.

Ohala, J. (1997, Aug). Sound symbolism. Proc. 4th Seoul International Conference on Linguistics [SICOL], Seoul, South Korea. Retrieved from

Sapir, E. (1929). A study in phonetic symbolism. [Abstract].Journal of Experimental Psychology,12(3). doi:. Retrieved from

Appendix I:

1.Brand Type: bicycle

Attribute: faster

a. /titən/

b. /tutən/

c. /tɑtən/

d. /tætən/

2.Brand type: Cell phone

Attribute: smaller

a. /kɑtɚ/

b. /kætɚ/

c. /kitɚ/

d. /kutɚ/

3.Brand type: blanket

Attribute: warmer

a. /pitəm/

b. /pɑtəm/

c. /pætəm/

d. /putəm/

4.Brand type: Ketchup

Attribute: thicker

a. /pukəb/

b. /pɑkəb/

c. /pækəb/

d. /pikəb/

5.Brand Type: Paper towel

Attribute: stronger

a. /tækəd/

b. /tikəd/

c. /tukəd/

d. /tɑkəd/

6.Brand type:jewelry

Attribute:prettier

a. /kupəl/

b. /kɑpəl/

c. /kipəl/

d. /kæpəl/

Stimuli

7.Brand Type: teddy bear

Attribute: friendlier

a. /tæpən/

b. /tupən/

c. /tipən/

d. /tɑpən/

Let the Writer Write and the Reader Read: The Infulence of Ourlipo on Computer-Generated Poetry, Erien Greenhalgh

(Back to top)

In 1984, free-lance writer William Chamberlain and program analyst Thomas Etter developed Racter, short for raconteur, a computer program that developed poetry (Ledbetter 39). In the same year, Racter’s creations were compiled into the first book of computer poetry entitledThe Policeman’s Beard is Half Constructed.This book includes poetry, prose, dialogue, and even conversations with its programmers. Though novel for its time, the beginnings of computer-generated poetry actually took form several decades earlier in 1950’s and 60’s France when the literary method of Oulipo was born. Begun by Raymond Queneau and Francois Le Lionnais, Oulipo, short for “Ouvroir de Litterature Potentielle,” stands counter to a traditional and romanticized view of literature that emphasizes personal inspiration and authorial genius; instead Oulipo embodies a systematic, mechanical, and algorithmic approach to writing, creating interesting works by highlighting the tension between structure and aberration. Oulipo, with its use of formal constraints, focus on literature in its potential state, and playful nature, parallels many aspects of Racter’s poetry, and as such becomes a useful tool to help us read and analyze Racter’s and other computer generated poetry. Readers of Oulipo also face similar frustrations as readers of Racter in that Oulipian works de-emphasize the importance of authorial intent. Reading Racter in the light of Oulipo helps us redefine the author not as one solely responsible for the work, but as one who assembles and sets the work in motion, freeing readers from the idea that an author ultimately controls a work’s meaning.

Oulipo and Racter’s poetry pose similar problems for readers because the works are separated from an author and the human mind. When readers confront a conventional work, they expect the text to make sense because they assume an educated and thinking author deliberately crafted the language. If the work does not immediately make sense, as in the case of much figurative and poetic language, the reader will work to understand it based on this same assumption of authorial intent. This need for an intending author is so strong that readers’ first instinct upon reading Racter’s work is to assume that the computer must somehow be thinking. InThe Policeman’s Beard is Half Constructed,readers will come across lines like, “More than iron, more than lead, more than gold I need electricity. / I need it more than I need lamb or pork or lettuce or cucumber. / I need it for my dreams” (Racter). Such lines are troubling because the computer refers to itself as “I”, and the work as a whole is self-reflexive. The computer seems to be aware that, unlike humans who need food to live, it needs electricity in order to function and to feed its “dreams,” which could be interpreted as these very lines or Racter’s poetic ability in general. When readers learn that Racter is merely a computer program and cannot think at all, the value of Racter’s poetry comes into question. They feel duped into believing the lines have meaning, when no meaning was consciously intended by a human author.

Similarly, Oulipian writings trouble readers’ expectations about the role of the author in relation to their work. Consider a key Oulipian work by Raymond Queneau titled “Cent Mille Milliards de Poemes” or “One Hundred Million Poems.” For this work, Queneau wrote ten sonnets in which the reader can substitute any line of the same number for any other line. The final result is 1014possible sonnets (Motte 3). According to Queneau’s own calculations, it would take the average reader over one million centuries to read the entire work (Motte 3). If no one reader (or even several readers) could discover all possible sonnets, obviously Queneau himself could not have either. So while the poem begins with intention, it quickly degenerates into random combinations of lines. Critics of both Oulipo and computer-generated poetry cite this apparent lack of thought as problematic: “where certain individuals see an example of original, conscious, and lucid poetic innovation, others will see only empty acrobatics, pretension, and literary madness” (Motte 3). Skeptics argue that Racter and Queneau are only literary magicians, manipulating symbols to simulate writing that has no literary value.

In order to appreciate computer-generated poetry and Oulipo; however, it is necessary to understand method and algorithm as a type of thought, an intention in itself. This symbol manipulation is not “empty acrobatics”, but a conscious creative choice, (perhaps the only creative choice that authors are free to make.) One hallmark of Oulipo is its use of formal constraints. Oulipian writers work within strict boundaries, relying on devices like lipograms, palindromes, or in Queneau’s case, the sonnet (Motte 11). The belief is that strict adherence to a constraint allows a greater freedom for aberrations or problems in the structure to surface and add interest to the work. In this way, highly structured Oulipian works anticipate formulaic computer poems. Like a dedicated Oulipian writer, Racter can only create poems according to the rules of its design. As journalist James Ledbetter explains, Racter is programmed to “conjugate regular and irregular verbs, provide the correct antecedents for nouns and pronouns, and determine the form of its own output” (Ledbetter 39). By generating and regenerating phrases of text, Racter creates poems with coherent subjects and themes (Ledbetter 39). The explicit use of rigid constraints in both Oulipo and computer generated poetry takes the emphasis off the author’s personal inspiration. As Queneau explains, “The classical playwright who writes his tragedies observing a certain number of familiar rules is freer than the poet who writes that which comes into his head and is a slave of other rules of which he is ignorant” (qtd. in Motte 18). The point is that all authors write under some type of constraint whether they realize it or not. Personal genius is an illusion. The real skill comes in mastering the use of a constraint to create interesting works, as both Queneau and Racter have succeeded in doing.

A second characteristic of Oulipo and Racter’s poetry is potentiality, a concept which positions the author as one who does not have ultimate control over the text. Oulipo as “potential literature” is interested in writing that could be, in the echoes of unwritten words that written words cause. Queneau’s “Cent Mille Milliards de Poemes” is a prime example of this concept. This poem consists of only ten written poems but of 1014poems that could be. Queneau himself does not control the poems thatcanexist. He only controls the poems thatdoexist, the original ten. However, his writing of ten poems in a strict formula allows almost infinite other works to be available, should one go to the trouble of arranging them. It is important to emphasize that potentiality is not the same as arbitrariness. Oulipian Claude Berge put it nicely: “Potentiality is uncertain, but not a matter of chance. We know perfectly well everything that can happen, but we don’t know whether it will happen” (qtd. in Motte 17). On one hand, the author is fully in control of the fact that potentials can exist. On the other hand, the author is not in control of those potentials. This holds true for Racter’s poems as well. By programming Racter, Chamberlain and Etter have set up a system in which potential works can exist. Though the way Racter writes is controlled, it is impossible to predict whether a certain phrase will occur or what text will be generated at all. So while Racter’s poems may appear to be random, they are rather the result of consciously intended potentials coming into existence.

Oulipo and computer-generated poetry also share distinct qualities of playfulness in which the relationship of the writer and reader to the work is fluid. As Warren Motte points out in his introduction to Oulipian tenants, “The Oulipian text is quite explicitly offered as a game, as a system of ludic exchange between author and reader” (20). Queneau’s poem becomes rather like a giant puzzle, daring the reader to take on the challenge of reading. Computer-generated poetry also includes play in the form of blazon confrontations from the text and semantic twists. On the back cover ofThe Policeman’s Beard is Half Constructed, Racter states clearly, “This book is about delight and satisfaction and joy…This is because champagne, seltzer, happiness, commitment, conflict, and tragedy are all in this volume” (Racter). In these lines, the challenge is not to construct poems out of potentials, as it is with Queneau’s work, but to face the assembled potentials, to face words greatly disconnected from human agency. In both Oulipo and Racter’s poetry, the challenge is simply, “Read.” Meaning from the text is not one-sided. That is, it demands participation from the reader in order to have full meaning. Motte titles his section about Oulipian play “Scriptor Ludens, Lector Ludens,” an apt title for a fluid relationship between writer and reader (20). The parallelism in the phrase, “the writer playing, the reader playing” makes it evident that it is the duty of both writer and reader to explore the puzzle of the text. Once readers acknowledge that they too are responsible for creating meaning, the anxiety of the issue of authorial intent lessens. The writer only assembles the work. It is the reader’s job to interpret it.

In conjunction with Oulipo’s formal constraints, potentiality, and playfulness arises the notion of the author as crafter, not creator, a distinction that is key for reading computer-generated poetry. “Ouvroir de Litterature Potentielle,” from which the abbreviation Oulipo is formed, translates to “the workroom of potential literature” (Motte 9). “Ouvroir” carries connotations of craftsmanship and hobby. Indeed, the word is used in French to describe a room where nuns work or where well-to-do women hold their sewing circles (Motte 9). The choice of “ouvroir” over a loftier term is a deliberate attempt by Oulipians to define their literature as the work of the artisan, separate from artistic genius. Canadian poet and critic Christian Bok explains Oulipian writing as “automatic, insofar as it results not from an aleatory impulse…but from a mandatory purpose. Writing is itself a machine to be studied methodically and guided systematically, as if by science” (Bok 64). In the case of computer-generated poetry, it is easy to understand writing as the “mandatory purpose” of the computer program. To understand the role of the author of these works, we can substitute the word “programmer” for “artisan.” A programmer assembles pieces of code, and forms the framework of the text, while the computer fills in the words. The programmer guides the shape of the work but is distanced from it; thus the work can take on meanings independent of an author. Though traditional authors work much closer to their texts, envisioning them too as programmers helps rid us of the notion that the meaning of a text is fixed in its author. In Ledbetter’s words, the author of any work “is no more responsible for the meaning of the message he produces than he is for the rules, or the fact that he is a coder…his existence is dependent upon the manipulation of symbols over which he has no control” (41). By making the role of the author as coder explicit, Oulipo and Racter’s poetry deconstruct traditional ideas about the source of meaning in a text.

Thinking of an author as an artisan or programmer allows us to read any work of literature, computer-generated or otherwise, with more sensitivity to the many layers of meaning available. When we critique a work based on the intent of the author who wrote it, we assume that all meaning in the work is tied to the author, and that the author is responsible for all meaning produced. We also create a persona for the author that does not necessarily reflect their actual self. We can create an idea of Queneau by reading his body of work and learning about his life and philosophies, but even with this researched picture we can’t be sure of what he “intended” when he wrote his ten sonnets. In the case of the 1014potential sonnets, we are less able to tie Queneau’s intention to the work. When we consider Racter, the “author” is even more removed, existing only as a programmer who set the computer in motion. There is no human mind to directly connect to the words; therefore, there is no authorial persona to construct. As Ledbetter points out, “by continuing to point to authorial intention as the origin of meaning, we conceal the fact that production of meaning is the project of the task of reading” (41). In fact, the very creation of an author-persona points to the fact that the meaning of a text is centered in the reader, not the writer. This persona does not exist except in the imagination of the reader. To elaborate on Motte’s Latin, we can think of making meaning in this way: scribat scriptor, legat lector—let the writer write and the reader read. The reader must recognize that to tie all responsibility for meaning to the author is to shirk their own responsibility. They must engage with the text, they must play the game, arrange the puzzle. This parallel responsibility between writer and reader exists for all works, but Oulipo and computer-generated poetry make this relationship extreme and explicit. By distancing the role of the author, Racter’s poems force the reader “to confront the processes by which they derive meaning” (Ledbetter 40). The reader recognizes Racter’s poems as literature not because they are filled with authorial inspiration, but because the poems hold the potential for meaning through formulaic structure and semantic playfulness. Once we become comfortable with the idea that the author as a programmer simply creates a space for meaning to exist, we can read all works of literature with more verve. When we encounter the lines, “A tree or shrub can grow and bloom. I am always the same. But I am clever,” we need not be afraid to throw ourselves into the puzzle and the challenge. We will say, “So you are, Racter.” And we will keep reading.

Works Cited

Bok, Christian. "French Oulipianism: A 'Pataphysics of Mathetic Expression."'Pataphysics: The Poetics of an Imaginary Science. Evanston: Northwestern University Press, 2002. 64-80. Print.

Ledbetter, James. (1986,August). Racter, the poetic computer: The case of the disappearing author.The New Republic (pre-1988),195(006),39. Retrieved November 3, 2011, from ProQuest Central. (Document ID:45745321).

Motte, Warren F.Oulipo: A Primer of Potential Literature. Normal: Dalkey Archive Press, 1998. Print.

Racter.The Policeman's Beard is Half Constructed. New York: Warner Books, 1984. Print.

(Back to top)