The Taste Manifold

The casual observer would probably feel that any attempt to enumerate and arrange in a logical scheme the infinitude of tastes and flavors would be an impossible task. To him it might seem that nearly everything in the world possessed its own peculiar taste. Such an observer would also be likely to think it impossible and thankless to attempt to reduce to their necessary limits the various kinds of substance of which this infinitude of things is made up. But the chemist would readily be able to show him that the infinitude of substances consisted, as a matter of fact, only of various forms and combinations of less than one hundred “elements,” and that from these elements one could produce, by appropriate selection and apportionment, any one of the infinitude of substances.

Is it then possible, in the field of our sensations, to reduce to elemental categories or units the manifold^[1] of concrete sense experience? In the case of visual sensations almost everybody knows that there are certain so-called “primary” colors, from which can be produced the whole range of color experiences known to man. Blue and yellow, red and green, these are the primary colors, and if to these, in their varying intensities, be added gray, with its range of brightnesses, we have the elementary components of all our visual experience. Such a distinctive color as that of fire clay, for example, may thus be said to contain, in specified degree and proportion, red, yellow, and gray, while the familiar color of a wild flower may contain, in specified relations, red, blue, and gray.

In a strict psychological sense, it remains true that each color experience is relatively unique and distinct. But it can readily be shown that these psychological fusions and compounds are elaborations of more unitary experiences which have as their basis distinct mechanisms in the nervous system and sense organs. For example, the sensation of “heat” is a readily recognizable and identifiable experience, yet the physiologist tells us that there is no separate sensory apparatus for this impression. Cold and warmth, we are told, depend on the stimulation of specific nerve endings. When these two types of endings, in the same general region of the skin, are simultaneously stimulated, as the result of the application of a stimulus with very high temperature, there arises that new experience of “hot,” which is in this sense a combination of warmth and cold.

Is it similarly possible to reduce to elementary units the rich manifold of taste and flavor? If this can be done, in what way must such an analysis proceed? What principles of classification are revealed, and what and how numerous are the elementary taste qualities? The various attempts that have been made to analyze the taste manifold are as interesting as their results are instructive.

The Classification of Tastes

One method of classifying sense qualities that has often been advocated uses as its basis the varieties of objects, agencies, or stimuli^[2] by the application of which the sense qualities are produced or aroused. Thus the whole field of sense experience might be divided into thermal, electrical, mechanical, photic (produced by light), etc. But such sensations as are aroused by electrical stimuli, for example, may be auditory, visual, cutaneous, gustatory (having to do with taste), etc., while these same varieties of sensory experiences may be aroused in some cases by mechanical stimulation. Hence the classification of stimuli does not yield an adequate analysis of modes of sensation. In the field of taste this method, although it has been seriously attempted, is equally futile. Thus various writers have attempted to group taste sensations according to the species of plants and animals whose tissues possessed sapid (taste-producing) qualities. It is obvious that this method is unsatisfactory, since it is by no means true that all specimens of a given vegetable taste alike. Even different parts of the same plant have widely different tastes, and indeed the taste of a given part varies widely with time and circumstances.

Even Chevreul, a famous early student of the sense of taste, adopted a chemical classification, on the basis of the composition of the substance tasted. Here again it is true that substances chemically very dissimilar may possess tastes which are strikingly alike. Thus acetate of lead, chloroform, and cane sugar, which, chemically considered, are very dissimilar, may easily be mistaken for each other if their taste alone is relied on; while starch, which is chemically closely related to sugar, has no taste. It is also true that such different tastes as sweet and bitter may characterize substances which are chemically very closely related.

It is, however, true that certain broad lines of chemical classification may be drawn. Thus those substances belonging to the colloid^[3] group are tasteless, the crystalloids all being sapid. When substances are arranged according to the “periodic law” of chemistry, the elements present in sweet-tasting substances are in general neither extremely positive nor extremely negative. While it is the general rule that soluble alkaloids are bitter, acids sour, sugars sweet, and salts salty, there are many curious exceptions in every case. “It is true that we get the taste of salt only from chemical salt; but there are chemical salts that taste sweet, others that taste bitter, others again that have no taste at all.” Similarly, while it is true that sour tastes belong to acids, it is by no means true that all acids taste sour. Moreover, sugar of lead, which is a salt, tastes sweet; while sulphate of magnesia and other salts taste bitter.

Indeed, it is true that there are substances which have more than one taste, the taste varying with the region of the tongue at which the substance is applied. Thus saccharine, sulphate of magnesia, and acetate of potassium are said to have sweet or acid taste if applied to the side or tip of the tongue; whereas they are bitter if applied to the posterior part. There are various other substances which show similar changes in taste according to the point of their application. However such facts may be explained, it is clear that the classification of taste along chemical lines is not only beset with difficulties, but that even in attempting such classification we resort to the use of a more immediate classification, indicated by such words as sweet, sour, etc.

This resort to an immediately descriptive classification suggests that the various taste experiences, regardless of the stimulus provoking them, have certain similarities as direct experiences. This further suggests that a strict psychological classification, based on the attributes of tastes themselves, should be found through analysis. In the case of sensations in general, such a type of classification is the one that seems most satisfactory. Certain sense experiences, such as red, yellow, orange, seem, as a matter of immediate experience, to belong together and to be essentially different from such experiences as warmth, tickle, noise, dizziness, etc. Furthermore, it is found possible to pass by gradual steps of transition from red to yellow, through an intervening orange, while there exists no such intermediate region between red and tickle. As a matter of immediate experience, then, and regardless of the nature of the stimulus, or, so far as we may be aware, of the part of the body stimulated, certain sense experiences seem to belong together, to constitute a certain mode of sensation, such as pressure, sound, etc.

Is it now possible to apply a similar test to the various qualities which comprise the mode or sense of taste, and thus arrive at an adequate classification and analysis of these qualities? The earliest attempts to analyze the tastes by this psychological method were often amusingly miscalculated. Thus Chatin, in 1880, presented a scheme in which the total manifold of taste was first divided into agreeable tastes and disagreeable tastes. The agreeable tastes were typified by those we call sweet, and the disagreeable by those we call bitter. It was, of course, at once necessary to indicate certain intermediate conditions in this scheme for a variety of tastes which were neither clearly agreeable nor markedly unpleasant. Moreover, it is a matter of common experience that a taste which is agreeable to one person (such as tobacco, olives, mustard) may be decidedly obnoxious to another person, or, indeed, even to the same person on a different occasion; so that such a classification cannot be said to represent in any fundamental way an analysis of tastes.

There have been a great variety of classifications proposed on this direct descriptive basis, and a comparison of the various schemes at once suggests that the task is by no means as simple as it might seem. The number of elementary tastes ranges widely, some investigators enumerating five or six times as many fundamental taste qualities as others have recognized.

Haller enumerated twelve different qualities—stale, sweet, bitter, sour, sharp, tart, spicy, salt, urinous, putrid, spirituous, nauseating. It is evident that this classification represents only a transition step toward a psychological analysis and that it is by no means free from the suggestion of provoking substances (spirituous, putrid) and the suggestion of effects produced (nauseating).

LinnÆus recognized somewhat fewer categories,—giving the following ten as fundamental,—sweet, spicy, oily, mucous, salt, styptic,^[4] bitter, sour, aqueous, and dry.

Other authors have been content with indicating eight elementary tastes. Both Bain and Wundt have proposed a sixfold classification, as follows,—sweet, bitter, saline, alkaline, acid, and astringent or metallic. Most modern authorities reduce the number of elementary tastes to four,—sweet, salt, sour, bitter,—while at least three investigators have advocated a simple twofold classification, into sweet and bitter.

These divergent accounts of the elementary taste qualities are in large measure to be explained by the exceeding complexity of those experiences which we in everyday life refer to as “tastes.” It was long ago shown that a classification of the various senses on the basis of the gross “sense organs” or parts of the body involved is as inadequate as one based on the nature of the stimulating agent. The eye as a gross sense organ yields experiences of pressure, pain, temperature, and strain, as well as experiences of color and brightness. But these varied sensations we recognize as belonging, as experiences, to quite distinct modes. Even more complex are the varied sense experiences which we may receive through stimulation of the tongue and the surrounding tissues. For the tongue as an organ yields not only sensations of pure taste quality, whatever these may be, but it also gives rise to experiences of pressure with the varying characteristics of smooth, rough, moist, dry, contact, tickle, etc.; to experiences of pain, with the ranging characteristics and intensities, such as sting, smart, prick, burn; to experiences of temperature, such as cold, warmth, heat; and to a vast complex of kinÆsthetic or muscular experiences of contraction, torsion, strain, expansion.

In common experience these qualities of pressure, pain, temperature, and kinÆsthesis are scarcely discriminated from the purely gustatory or taste qualities themselves. Thus we speak of “oily,” “fatty,” and “greasy” tastes, in which the “smoothness” is certainly identical with that felt by the fingers and other parts of the skin. Similarly the pungent, astringent, puckering, biting “tastes” may come from substances which have no taste at all in a strict sense, but which produce definite smarting or stinging sensations when applied to the surface of the skin or to exposed nerve endings (pepper, camphor). The puckering quality can be shown by the characteristic muscular reaction to be largely kinÆsthetic and tactual in its origin. The difference in “taste” between cold ice cream and the same substance when melted indicates how much of the flavor is due to touch sensations and sensations of temperature. One observer, indeed, reports the experience of four different qualities of sensation from the stimulation of a single papilla,—a touch, a temperature, a taste, and a pain sensation.

Further, many substances, in addition to these locally aroused experiences of touch, temperature, pain, and movement, set up strong organic reactions in more or less remote regions as well as strong affective reactions: such as choking, nausea, and vomiting, on the one hand, and extreme unpleasantness, disgust, distress, strain, and shock, on the other. In many cases these immediate reactions seem to be reflex or instinctive in their origin, and in other cases they seem to be conditioned reactions, based on past experiences and associations. Thus in one case the “taste” of ice cream, which was once agreeable enough, has come to be immediately nauseating in character.

In spite of all these facts many of the classifications of taste qualities have included the “oily,” the “nauseous,” the “astringent,” etc., as primary taste experiences. Even if the complications we have thus far alluded to were the only ones concerned, it would be clear enough that the “tastes” and “flavors” of everyday conversation represent very complex fusions and compounds, and that an analysis of the true taste qualities, if such there be, must take these factors into account in some very careful experimental way.

But we have left until this point a single complicating factor which in itself is sufficiently serious to call for very careful technical procedure in the examination of the sense of taste. This is the fact that a very great number of our so-called tastes are not tastes at all, but really odors. The sense organ of smell is so situated that it may be stimulated not only in the ordinary way, through particles borne into the nostrils by currents of air from the outside, but also by particles and vapors which pass up, from the mouth cavity, behind the soft palate, by way of passages called the “posterior nares.”

In this way it happens that tasteless substances, with definite odor, are mistakenly supposed to have taste. In 1824 Chevreul reported a very simple experiment with which his name has since been universally associated. He pointed out that it is impossible to separate the action of a substance on the touch corpuscles of the tongue from its action on the taste buds themselves. He observed, however, that by a very simple expedient it is possible to eliminate to some degree the factor of odor. His classical experiment consisted in excluding the sense of smell in large part by pressing the nostrils with the fingers while the substance to be examined was presented to the tongue. In this manner he observed that a piece of camphor gum which had seemed to have a very distinctive taste had in reality no taste at all. When the nostrils were closed all that could be observed as the result of placing camphor on the tongue was a peculiar pricking sensation of touch, similar to that produced by various other substances. The sensation produced by the camphor was thus not a taste at all, but a fusion of odor and touch.

If, under the simple conditions of Chevreul’s experiment, the various substances be reduced to a state of like consistency, so that they cannot be recognized by the tactile sense, observers are usually much amazed to discover that through taste alone it is impossible to distinguish between quinine and coffee or between apple and onion. Many familiar experiences of daily life testify to the large contribution which the sense of smell makes to the supposed taste. How “tasteless” are our fruits, wines, cigars, and vegetables when one has a cold in the head, and the free passage of odorous particles to the organ of smell is obstructed! How often has the nasty taste of medicine been softened by Chevreul’s simple technic of “holding the nose”! There are some cases in which the reverse of this situation occurs and volatile substances, entering the mouth through the nostrils, stimulate the taste buds in the upper and back part of the mouth. In such relatively rare cases the real taste is mistakenly interpreted as an odor. In this way chloroform seems to have the characteristic odor which is in all probability a sweet taste due to stimulation of the taste buds by the chloroform vapor.

Why should it be the rule that, since the taste and smell qualities are to be confused, smell should so commonly sacrifice its claim, so that odors are called tastes rather than vice versa? No doubt this is true largely because of the customary presence of sensations of pressure, temperature, movement, and resistance which are localized in the mouth and in the organ of taste. These accompanying sensations suggest that the taste organs are active in determining the result, even when no true taste qualities are present. It is common for sensations to be displaced in some such way as this, just as the blind man, who is really getting sensations from stimuli in the palm of his hand, seems to be getting them at the end of his walking stick; or just as a faint sound may seem to come from any source to which we direct our attention. Similarly, the whole complex of touch, taste, and odor experienced as the result of “sniffing” at a particular substance are quite likely to be credited entirely to the sense of smell.

The Poverty of Taste

Here, then, is a most interesting situation, which has been described by the use of two apt phrases: “the poverty of taste” and “the self-sacrificing character of smell.” Our analysis has tended constantly to rob the sense of taste of the richness which we ordinarily credit to it. In fact, modern authorities agree that there are only four qualities which can be truthfully classed as tastes, namely, sweet, salt, sour, and bitter. What we took to be the taste manifold is really a meager equipment of four qualities, with such variations of intensity and combination as these may possess. Both the richness and the manifoldness come from qualities of other senses, parading under the guise of taste. Smell especially is prone to sacrifice its claim in favor of its neighbor, and it is common indeed for us even to use taste names in describing odors; we speak of “sweet odors” and “sour smells,” although sweet and sour are primarily taste qualities. Smell, then, not only entirely yields many of its qualities to the various “taste blends,” but allows some of its own sensations to pass under taste names.

Patrick has reported extended observations in which he studied the taste experiences of an anosmic,—a person who had lost the sense of smell. In some of his experiments this woman, with two other women as control subjects, after having been blindfolded, attempted to identify various substances taken into the mouth. The general principles on which the experiments were based are stated as follows: “In theory those substances not recognized by any of the observers depend for their recognition upon sensations of sight; those recognized by the normal observers but not by the anosmic depend upon sensations of smell for their recognition; while those recognized by all observers depend upon either taste, touch, or muscle sensations.” It was further suggested that those substances recognized by the anosmic but not by the normal subjects would seem to depend in the main upon touch or muscle sensations.

These experiments disclosed many curious and unexpected facts. Breads and meats, butter, cream, olive oil, and various fruits and vegetables could not be easily identified when only sight was excluded. One of the women, a housekeeper of long experience, could not recognize raw turnip, raw potato, boiled pumpkin, cranberry sauce, or fresh pear when she was blindfolded.

Chicken, turkey, and quail were found to differ surprisingly little in actual taste, especially if their characteristic texture, smoothness, and other tactual qualities were eliminated or disguised. The various values placed upon different meats, breads, etc., in the general esteem would seem to depend in great measure on associated ideas and emotions, rather than on their actual qualities for taste.

Especially interesting is the list of substances which were recognized and correctly named by both of the normal women in these experiments, but which the anosmic was unable to identify. Patrick enumerates twenty-seven such common substances. Among them, by way of example, were vanilla extract, pineapple syrup, banana, grape, quince, strawberry, tea, chocolate, sour milk, kerosene, claret, rhubarb, onion, eggs, and boiled turnips. The results suggest that these substances, although they seem to have very characteristic tastes, are actually differentiated and recognized on the basis of their olfactory rather than their gustatory or tactual qualities.

One justification of this olfactory sacrifice is suggested by the fact that biologically one of the most important functions performed by smell is that of aiding in the discrimination of food. Through smell the animals perceive at a distance a substance which may offer itself as possible food. Biologically, the immediate guide to the acceptance or refusal of food is the sense of taste. In so far as smell is in part a subordinate servant in this matter, and hence becomes easily associated with such reactions as “eating” or “not eating,” no injury is produced through the occasional confusion of the two modes.

We have thus reduced the rich manifold of taste to the qualities of sweet, salt, sour, and bitter,—four meager qualities as compared with the numerous unanalyzable qualities of various of the other sensory modes. We have now to show by what logic, through what technic, and on the basis of what evidence, we are compelled to grant to taste four qualities rather than two or twelve, and why the final grant consists of these particular four rather than others.

Psychological Analysis of the Taste Qualities

As Chatin long ago observed, “The three senses,—taste, touch, and smell,—are so intimately combined that they seem to refuse to yield themselves to minute analysis.” These associations seem to be even stronger than those between the various taste qualities, of which Ladd and Woodworth have remarked, “On the whole it appears as if the four tastes were rather isolated from each other, each representing almost an independent sense. There is much blending, to be sure, but the amount is apparently no greater between one taste and another than between tastes and odors.”

We may now fairly ask how these four qualities may be made to reveal their elementary and independent character, once we have eliminated the complicating factors introduced by the intrusion of qualities from other senses.

The first appeal is to common observation and experience, according to which the four taste qualities,—sweet, sour, salt, and bitter,—stand out as conspicuous classes within which may be placed a great variety of “taste blends.” Thus many substances, while having more or less distinctive flavors, resemble each other in that unanalyzable quality which we call sweetness. The only question here is whether or not we should also include qualities other than these four as ultimate and irreducible. If we refer back to the lists of tastes proposed, we find that many of them are such as can be shown to be analyzable into one or more of these four qualities, plus the intrusion of tactile, thermal, kinÆsthetic, and olfactory qualities. Such tastes as nauseating, aqueous, astringent, styptic, putrid, etc., are easily ruled out on this score alone. Introspectively, by simple experimental variation, or by casual observation, the complex character of many of these tastes is easily revealed, and when the non-gustatory components are eliminated the residue falls readily under one or other of our four qualities.

These reductions are borne out by definite experiment. The tactile (touch), thermal (temperature), and kinÆsthetic (movement) factors are kept constant and reduced to a minimum by applying minute amounts of various solutions to single papillÆ or very small regions of the tongue. Smell may be, under these conditions, in great measure eliminated by closing the nostrils with cotton or wax, and by letting the tongue be somewhat advanced beyond its usual position. When these conditions are observed, it is found that the main sense qualities experienced are those of salt, sour, sweet, and bitter, along with such touch sensations as may be unavoidable. Temperature may be eliminated by having the solutions maintained at the temperature to which the tongue is already adapted.

The evidence on this point is not absolutely consistent. Some observers, for example, feel impelled to add metallic and alkaline to the group, making six elementary qualities instead of four. Other observers,—most, in fact,—are persuaded that the metallic and alkaline qualities represent mixtures of the salt, sour, sweet, and bitter, along with unavoidable sensations of touch and smell. Thus, by a suitable mixture of strong solutions of salt and sweet substances, the alkaline taste may be very well produced. “It has been suggested that the metallic taste is due to the simultaneous development of salt and sour tastes. The failure to produce exact alkaline and metallic tastes synthetically is in part due to the difficulty of imitating the tactual and other sensations with which they are bound up.” Still other observers are convinced by careful elimination of smell sensation that the unique character of the alkaline and metallic qualities is really a question of odor.

By the application of specific drugs to the organ of taste further indications are secured that these four qualities, unanalyzable to introspection, also function in relative independence. Thus, the juice of gymnema leaves temporarily destroys the qualities of sweet and bitter, while sensitiveness to sour and salt remains unimpaired. “The true acid or sour taste may be separated from the astringent effect which accompanies it by painting the tongue several times with a five to ten per cent solution of cocaine. Cocaine first abolishes the sour taste, and after several minutes begins to abolish the astringent action of the acid solution. Later, the sour sensation begins to return, while the astringent effect is still in abeyance, so that the application of an acid solution at a certain stage during recovery enables the true taste character of sour to be differentiated.” It is also reported that when gymnemic acid and cocaine are applied to the tongue, the one abolishes the sweet and the other the bitter, thus leaving the two other tastes relatively unimpaired. Certain mixtures seem to paralyze both sweet and bitter, but the former sooner than the latter.

Distribution of the Taste Qualities

To these four elementary tastes we are not equally sensitive on all parts of the sense organ. Roughly speaking, sweet is best tasted at the tip of the tongue, and many forms of candy are prepared so as to allow as much as possible the employment of this part of the taste organ. Bitter, on the other hand, is best tasted at the back or root of the tongue, which explains why many substances do not taste bitter until we have swallowed them. The edges of the tongue are most sensitive to sour, while in adults the central area does not commonly yield taste qualities at all. In children, however, the taste buds extend not only over the whole surface of the tongue, but are also found in the walls of the cheek, the palate, and even on the larynx and epiglottis. Titchener has suggested that these facts may explain “the childish tendency to take big mouthfuls.”

Attempts have been made to determine whether these elementary taste qualities depend on separate taste buds or papillÆ. Experiments show it to be true that some points respond only to sweet, sour, etc. But there are others which yield two or three or even all four qualities, while some yield no taste at all. In one such experiment thirty-nine papillÆ, in a certain region, were separately stimulated by acid, sugar, salt, and quinine. Of these thirty-nine, thirty-one responded to salt, and the same number to sweet, twenty-nine to sour, twenty-one to bitter. Four yielded no taste at all, one responded only to bitter, and one to sweet. In another case of one hundred and twenty-five papillÆ examined by solutions of sugar, quinine, and tartaric acid, sixty gave sensations of all three qualities (sweet, sour, bitter), twelve gave both sweet and sour, twelve sour only, seven bitter and sour, four bitter and sweet, and three sweet only. None of them seemed to give a bitter quality alone and this seems to be the general rule.

It is of course difficult in these experiments to restrict the application of the stimulus to single taste buds or even to single papillÆ. But these experiments, along with the effects of drugs which we have already described, suggest that the taste buds are not all alike in function, even though they seem quite similar so far as appearance and structure are concerned.

The Vocabulary of Taste

Several investigators have been interested in the study of the taste names found in different languages and communities. It has been suggested that such a study might throw light on the order of development of the various taste qualities.

Kiesow found that both children and adults quite commonly confused bitter with salt and sour. Myers found, in studying the taste names of primitive people in the region of the Torres Straits, that they had no word for bitter. In some primitive languages the same word is used for sweet and salt. When there is a word for salt it is usually some derivative of the word for sea water. Salt and sour are also often confused. In Polynesia, indeed, a single word is used in describing salt, sour, and bitter tastes. This is analogous to the fact that in primitive languages it is often found that the same word is used to indicate blue and black.

Attempts to argue from these facts of vocabulary to facts of sensitiveness and order of development are, of course, open to many sources of error. As Myers points out, “The differences between sour and bitter are considered less striking than their common unpalatability.” It has often been pointed out that in our own language sweet is probably the only taste word that had from its very origin a gustatory meaning. In some languages even the word for sweet means literally “tasting good.”

Vocabularies do not develop in order that structural and functional facts may be recorded for the information of forthcoming scientists. Words arise in response to the demands of practical life. It is practically more important that some substances “taste good,” and others “taste bad,” than that there are just four elementary taste qualities. Hence for certain primitive circumstances two taste words are all that are needed in ordinary conversation. It by no means follows from this that the salt, sour, and bitter, which all fall in the “bad taste” category, are not discriminable from one another by the taste organs of the savage. It would be just as cogent to insist that, since we have only one word for the taste of various sour things, all of these various tastes must be indistinguishable to us.

Nor is the argument safe that those sense qualities for which specific names exist must be more ancient than those qualities for which names are borrowed. Many of our color names are not primarily color names at all,—as violet, rose, olive, turquoise, lemon, straw, orange, and, perhaps, pink and green. Red, blue, and yellow seem to be more essentially color names. Yet, it is difficult to suppose that an organism sensitive to red and yellow should not also be sensitive to orange, which may be produced by a mixture of red and yellow light.

In the case of the odors, which we have every reason to believe are extremely ancient sense qualities, we have in our own language almost no exclusively olfactory names. Smells are designated by the objects with which they are associated,—as lilac, lavender, musk; or names are borrowed from other sensory modes, as sweet, sour, heavy; or still more descriptive and perceptual names are used, such as fresh, flat, rancid, foul, nauseating. Interesting as the vagaries of vocabulary may be, they yield very little information concerning the primitiveness, elementariness, sensitiveness, or distribution of the various taste qualities.