The Porter Zone

Philosophical musings and more

The Gendered Gaze

Fifty years ago I put on pants and walked the middle road

Katharine Hepburn


This essay is by way of being a reaction to MaryAnn Johanson’s definition of a female gaze analogous to the ‘male gaze‘ of film theory.  Johanson’s approach is a valuable step towards a more generally applicable theory, but her current definition inherits from  ‘male gaze’ theory a somewhat essentialist view of human gender. The problem lies in the implicit assumption that people can be categorised as being ‘male’ or ‘female’. Indeed, it is not even clear what those terms mean. I will argue that there are at least two independent factors that contribute to a person’s appreciation of images of other people, and that both of these are insusceptible to rigid categorisation.  This means that in addition to the ‘male gaze’ and ‘female gaze’ (both of which are now approximations) there is an ‘in-between gaze’.

It could be argument that my approach (which takes in Jungian depth psychology, probability theory, epistemology and some deftly hidden analytic topology) is using an intellectual sledgehammer to smash a nut whose contents are well-known.  Maybe so.  But I, at least, find it satisfying to understand precisely what are the assumptions on which such obvious conclusions rest, and to see the steps in the argument leading to them set out with great care.

Note that my purpose in this essay is to examine the definitional issues relating to gendered gaze theories.  I do not discuss the function of the gaze in defining gender roles, partly because that is a very complex topic, partly because until we actually know what the gendered gaze is such a discussion is pointless.  Therefore this essay attempts to place the concept of ‘gaze’ on a sound footing, as a preliminary to an analysis of its wider cultural ramifications.


I state my conclusions here, for those who are prepared to take the argument on trust.

  1. When treating people en masse, we can speak only of instinctive preferences, as opposed to intellectually motivated preferences. So I might really be homosexual, but pretend to be heterosexual because of societal pressure. Intellectually motivated preferences are too complex for our discussion, as they have to be treated on a case-by-case basis, whereas instinctive preferences, being simpler, are susceptible to generalised arguments.

  2. Gender and sexuality cannot be treated as simple dichotomies. People are predominantly-homosexual, predominantly-heterosexual or of mixed orientation, while being predominantly-masculine, predominantly-feminine or of mixed gender. No combination of gender and sexuality is forbidden (though not all  need be equally likely).

  3. Sexual attraction is based on an individual’s gender and sexuality, but not their sex. This is, on consideration, obvious, because what matters is not what sex one is, but what sex one thinks one is. It leads to some rather startling conclusions, such as that a feminine homosexual man will prefer women as sex-partners, and so appear like a masculine heterosexual man.

  4. We can derive a precise quantitative probabilistic model for an individual’s sexual preference, based on their gender and sexuality. In terms of qualitative models, we can choose between a nine-fold typology, a five-fold typology or a three-fold typology, depending on how much detail we want to capture from the underlying quantitative model. Again, classification into a simple ‘male’ / ‘female’ dichotomy is impossible.

  5. For our purposes the three-fold typology is sufficient: predominantly man-preferring (which includes, but is not limited to, the traditional ‘female’ category), predominantly woman-preferring (which includes, but is not limited to, the traditional ‘male’ category) and mixed-preference (which is new). The sexual reaction to images of other people will be reasonably predictable for individuals in the first two classes (so if I was predominantly woman-preferring, I would respond appreciatively to images of beautiful women and only minimally to images of beautiful men). The interesting class is mixed-preference: the reactions of people in this class are essentially unpredictable.

  6. Aesthetic preferences may be more or less correlated with sexual preferences, but remain distinct because they are influenced by culture and learned taste rather than instinctive reactions. Therefore there is little of a general nature to be about aesthetic preferences.

  7. Given that all of us (barring a vanishingly small number of extreme cases) are constituted of both male and female parts, though it is not true that one sex’s body is more desirable than the other’s, it is plausible that there should exist universal standards for what make a man or woman attractive (in Jungian terms, these would exist within the collective unconscious).

A note on terminology

Throughout this essay I shall use the words man and woman and their adjectival forms manly and womanly to denote the biological sexes. So I take ‘man’ to denote an individual with a Y-chromosome, and ‘woman’ to denote an individual without. I am aware that there are a small number of confusing cases where biological sex is unclear: double-X men, XY-women, hermaphrodites and other liminal cases; however, this is a complication that adds little to the argument, so let us make, for now, this simplifying assumption (which, after all, works over 90% of the time).

I shall use the words male and female and their adjectival forms masculine and feminine to denote genders, which reflect the psychology of the individual rather than their anatomy. I shall argue below that gender and sex are essentially independent, so it is perfectly possible to be a feminine man, for example (this is what I would expect to be the initial state of a man-to-woman transexual).

I shall use the words homosexual and heterosexual with their familiar formal meanings, with the attraction based on sex and not gender. So a homosexual is one who feels sexual desire for individuals of the same sex, while a heterosexual is one who feels sexual desire for individuals of the other sex. So sexuality is taken as being about attraction to types of body, not societal roles. I will (apart from within this sentence) not use the word bisexual, as part of my argument will render it essentially unnecessary.

In contexts where it is clear that some kind of sex/gender-related meaning is intended, but it is not clear what the explicit intention of the originator of the concept under discussion was, I shall use the words ‘male’ and ‘female’ with scare-quotes. I apologise if this becomes tedious, but unfortunately we have too few words with which to describe complex concepts (after all, the mere idea that sex and gender are not identical is relatively recent), and precision is a crucial part of my argument.

Finally, when I say two attributes of a collection of things are independent I mean that all possible combinations of values of the two attributes occur in the collection: there are no ‘forbidden’ combinations. This does not mean that there is no correlation between the attributes, in the sense that certain combinations may be more or less common. I am concerned not with how oftencombinations occur, but with whether they can occur at all.

Possibility and actuality

In the definition of independence I said I was concerned with whether states can occur, not how often they occur. In my analysis of sex, gender and sexuality I will be concerned with possibility: what combinations of these variables can occur and which are forbidden. So when I say that a particular value or combination can exist I am not saying I can point to a thing with that combination of attributes. I am saying that there is a non-zero probability of such a thing existing; its existence is not forbidden.

Thus, later on I will demonstrate that all possible gender mixes between the two extremes of pure masculinity and pure femininity can exist, and similarly for sexuality, and for combinations of gender and sexuality. This does not mean that there are living today, or have ever lived, exemplars of each possible mix or combination, but rather that there is no mechanism inherent in human biology or psychology that prevents such a mix or combination from being realised. This point is absolutely crucial, as it allows the use of thought experiments to consider possible cases that, so long as they are psychologically plausible, can be said to be realisable, if not realised.

Understanding the problem

Gendered gaze theory

The original theory of the ‘male gaze’ is an attempt to explain the over-representation of women in visual art. The thesis is that essentially this is because prior to the twentieth century the majority of art patrons and artists were men, and that though some of them may well have been homosexual, societal disapproval would force them to give the appearance of being heterosexual. Heterosexual men prefer looking at women it is asserted (though I question whether this is automatically true), so we see many beautiful women in visual art, but few beautiful men (the obvious exception is in Carravagio’s work, but he is unusual as an artist in so many ways that he does not constitute a significant counterexample). Such well-known women artists as Elizabeth Vigee Lebrun painted beautiful women because by the time a woman could actually make a living as a painter, the convention that beauty was a property of the womanly body alone was deeply ingrained.

And so art-theorists who argue that it is true that womanly bodies really are more beautiful than manly, or somehow have aesthetic properties that those of men do not are reflecting societal pressure. Indeed, it is rather startling that in so subjective a field as aesthetics anyone would even begin to think they could assert that any position taken was more than opinion: the mere fact that some appear to argue for the aesthetic primacy of the womanly body as some form of natural law should be a warning sign that something strange is happening. And similarly, women who allow themselves to be convinced that beauty is theirs alone are reflecting centuries of indoctrination by a heterosexual-man-dominated society.

Thus the basic ‘male gaze’ theory. Johanson’s proposal is that, as we no longer (at least ostensibly) have a heterosexual-man-dominated society, the time has come to reclaim the beauty of the manly body, and so in addition to the woman-preferring ‘male gaze’, there should be a man-preferring ‘female gaze‘.

Problems with the theory

I have placed scare-quotes around the expressions ‘male gaze’ and ‘female gaze’ as it is not clear what the words ‘male’ and ‘female’ mean within this context. Some obvious problems are as follows:

  • Is the ‘male gaze’ the property of heterosexual men or heterosexual males? Or both? (and mutatis mutandis for the ‘female gaze’)

  • Do homosexual women count as being ‘male’ for the purposes of the theory? (and mutatis mutandis for ‘female’)

  • Is it really the case that each of us find pleasure in contemplation of a particular sex’s body based on being ‘male’ or ‘female’?

  • Is it really possible to produce a simple classification of people as ‘male’ or ‘female’?

Technical background material

The sorites paradox

The sorites paradox has been a rich source of philosophical debate for millennia, so I do not pretend to give an exhaustive discussion (no such could be possible). I shall merely state what I need.

The paradox is most simply explained through an example. Say I have a rectangular strip which is red at one end and blue at the other with a continuous modulation from one colour to the other in between. It might look like this:


I now have three facts:

  1. One end is red.

  2. The other end is blue.

  3. If I start at any point on the strip and move a very small but non-zero distance in either direction, the colour does not change (in terms of what we call it).

So we get a paradox: start at the red end. Step along the strip using steps small enough for fact 3 to be true. Eventually you reach the other end, so red = blue. Contradiction. So at some point fact 3 must have failed, but the problem is that any attempt to define precisely where that is results in nonsense, because wouldn’t the points just to the left or right do equally well, in which case we are off again, and get another contradiction.

This example actually repays further study. At each of the two ends of the strip there is a region in which the colour is clear: close enough to the left-hand end it is clearly red, and close enough to the right-hand end it is clearly blue. It’s what happens in between that’s interesting, because about half-way along the strip is purple: neither red nor blue. So there is no systematic way of saying whether the colour is red or blue: there is no hard and fast dividing line (which this example makes happily self-evident).

There are many ways of resolving this paradox. I will present the one that I find most convincing, because (a) it is conceptually simple, (b) it is highly intuitive, and (c) it is consistent with what I believe to be the correct resolution, but as that requires three-valued logic it would take us too far afield (those interested in the details should see Understanding Truth by Scott Soames). However some philosophers (e.g. Roy Sorensen) reject this resolution’s consequence that a sorites problem inevitably leads to radical unknowability.

The resolution is as follows. Imagine a group of people looking at the strip independently. Each of them sees an indisputably red region and an indisputably blue region. Moreover each can assign some point they consider to be the best choice for the dividing line between blue and red (or at least a region within which they believe such a point should lie). Now consider the consensus view. They all agree that a region near one end is definitely red, and another region near the other end is definitely blue. They will not agree on the position of the dividing line (or region). So, on consensus, there is an indisputably red region, an indisputably blue region and a debatable region where there is no systematic way of assigning a colour.

So after this argument we want to take away two basic ideas:

The sorites problem

Say we have a collection (finite or infinite) of things (points on the strip in the example) such that:

  • The things can be assigned a property within some range (so in our example this property is colour).

  • A way of assigning one of two kinds to a value in the property’s range, such that values at one end of the range are of the first kind, while those at the other end are of the second (so in the example, this looks at the colour of a point and says whether it is red [at one end] or blue [at the other]).

  • Sufficiently close values have the same kind; so things with sufficiently similar properties have the same kind (this is just fact 3 from above).

Call this a sorites problem. Given a sorites problem, entities can be categorised as follows:

  • For each of the two kinds there is a region of certainty, consisting of things to which we can definitely assign that kind.

  • All other things belong to a region of uncertainty, where there is no systematic way of assigning a kind; the best one can do is to assign kinds on a case-by-case basis.

So in the example, the parts of strip near the two ends are the regions of certainty for red and blue, while the purplish area in the middle is the region of uncertainty.

The strong sorites argument

Now suppose we have a sorites problem, such that:

  • The things in the population have a population of things with a property, and a way of classifying them into one of two kinds based on that property,

  • It is provable that small changes in the value of the property never change the classification,

so fact 3 never breaks down, as we forced it to in the discussion of the paradox above. Then, based on our initial discussion of the sorites paradox, it is easy to see that:

  • If one thing in the population is of a kind, then all things in the population are of that kind.

I call this the strong sorites argument.  The crucial difference from the sorites paradox is that resolution of the paradox forces fact 3 to break down, whereas now we are forcing it to be universally true.

It can be quite hard to find categorisations into kinds for which fact 3 never breaks down. In fact we will only use one categorisation: ‘can exist’ and ‘cannot exist’. Unlike the kinds that leads to sorites paradoxes, this knows no middle ground (ignoring some rather complex epistemological issues that would enormously complicate the argument without noticeably adding to it), so we can be confident in asserting that, for it, fact 3 is true everywhere.

Jungian gender theory

Jung’s analytic psychology is interesting in that he does not take an essential view of gender. So rather than asserting that, for example, all women are purely feminine, he asserts that each of us has both masculine and feminine components, the balance of which determines our gender. This is in marked contrast to the essential quality of sex and gender asserted by other psychological theories.

To simplify the discussion of the theory, I will assume that my subject is a woman. So a woman’s conscious mind is dominated by a female component. However, there is a key aspect of her psyche which is her male component: her animus. This is part of her subconscious mind, but depending on its development it may have a greater or lesser influence on her conscious mind, thus creating a greater or lesser masculine component of her personality. So a woman can range from being almost purely female (undeveloped, and hence wholly unconscious animus) to being predominantly male (overdeveloped animus). This whole discussion goes through mutatis mutandis for men, with sex and gender terms interchanged and replacing animus with anima.

Finally, a quick note on individuation. Individuation is the process of psychological integration that the psyche undergoes through an individual’s life, resulting in a fully developed personality, resulting in a distinct and well-defined individual. The development of the unconscious gender component referred to above is one part of this process.

Analysis 1 : sex, gender, sexuality

In this section I start the analysis by examining the notions of sex, gender and sexuality and their relations to one another. The key results are:

  • That rather than the simple dichotomies used in popular discourse – male, female; homosexual, heterosexual – both gender and sexuality are continua (in the sense that no state lying between the two extremes is impossible to realise), with some individuals being clearly defined as being of one kind or the other, while others are simply a mixture. So there are three types: predominantly at one extreme, predominantly at the other and mixed.

  • That sex, gender and sexuality are independent (in the sense defined above), so no possible combination is forbidden.

Psychological variables are continuous

My argument depends critically on one hypothesis:

  • If I can find an individual for whom a psychological variable (e.g. gender, sexuality) takes a specific value, then nearby values of the variable can be realised by individuals in the population (albeit possibly with very low probability).

(Recall that can implies possibility – that it is not impossible – rather than that such individuals actually exist). Rephrasing the hypothesis less formally: the ranges of psychological variables are (in some sense) ‘smeared’.

To see why this is a plausible assumption, say it is not true. If I know an individual who is assigned the value 90%, what could happen is either:

  1. Values on either side of 90% can occur, which is consistent with the hypothesis.

  2. The variable can take values close to and greater than 90%, but none in a range just below 90%. This means that there is a hard edge at 90%, so the entire population exists either above 90% or some way below it.

  3. The value at 90% is isolated, so individuals are either assigned the value 90%, or a value some non-zero distance away from 90%.

Cases 2 and 3 are those incompatible with the hypothesis. Such hard edges and precise values do not occur in psychology, where behaviour is never hard-edged or precise. Small changes in the structure of the psyche are endemic. As a thought experiment, I can perturb an individual’s psyche slightly to achieve the required small change in the selected variable. As the result is a possible human psyche, there must be a non-zero probability of it existing in the population.

Looking at the specific cases I am interested in, it is intuitively obvious that if an individual can be homosexual 90% of the time, then it is not impossible for an individual to be homosexual 89.9% of the time, because that requires only a tiny change in the individual’s behaviour. Similarly, if my anima is so developed that I an 70% feminine, then at some point it must have been slightly less developed, meaning that I would have been 69.9% feminine, and if I developed it a bit more then I could be 70.1% feminine.

Therefore the contradictory scenario is extremely implausible, which means that my hypothesis is extremely plausible. Call it the continuity hypothesis. It allows us to apply sorites-type arguments to gender and sexuality.

Gender and sexuality are not simple dichotomies

The method of attack is to show that gender and sexuality are sorites problems, so each results in a classification into two predominantly pure types and a mixed type.


It follows from Jungian gender theory as described above that an individual’s gender falls within a range, with fully-male at one end and fully-female at the other. An individual’s position on the range is determined by the degree of development of the unconscious gender component (I write this being, as I am, a strongly feminine man). Assigning an individual a label male or female and using the continuity hypothesis gives rise to a sorites problem, and so by applying the sorites paradox we get three gender classes: predominantly-malepredominantly-female and mixed-gender.


Unfortunately, there is no good equivalent to Jungian gender theory for sexuality, so I shall have to use a longer elementary argument. The existence of people of homosexual and heterosexual inclination is a given. And, though there seem to be trends in certain circles to adopt the essentialist view that one is either homosexual or heterosexual, and there can be nothing between (see, for example, Mann’s biography of Katharine Hepburn, which seems to argue that as Hepburn had very strong loving relationships with women, her well-documented relationships with men were somehow not real; see also attempts to label Virginia Woolf as a pure lesbian, which end up having to deny that she loved Leonard Woolf, rather contrary to observed fact), it is not hard to think of examples, numerous examples, of people who inhabit the region in between: predominant heterosexuals who nonetheless have homosexual affairs, predominant homosexuals who nonetheless have heterosexual affairs, and individuals who take part, to a greater or lesser extent, in both orientations.

Having established that sexuality is not a binary opposition, but that there exist states in between, concluding that there is a continuum of possible states is relatively simple. It would be nonsensical to assume that all the individuals who spend most of the time in one orientation spend exactly the same amount of time in the other, so we have a blurred region starting at 0% which contains people, and another ending at 100% homosexual. Likewise it would be nonsensical to assume that all ‘in between’ individuals are precisely 50% homosexual, or even that they all share same value, so there is another blurred region around 50%. So we have something like this (where orientation changes from homosexual to heterosexual as one moves across the page):

Now, one could model human sexuality as consisting essentially of polar opposites, with a certain amount of deviation due to psychological factors (experimentation, deliberate transgressive acts, simple human cussedness, etc), but is hard to find any plausible model that allows two polar opposites and a mixed region in the middle but nothing in between.

So, let us set up a sorites problem. At one end we have 0% homosexual, a neighbourhood of which we know is populated. At the other end we have 100% homosexual, a neighbourhood of which we know is populated. Now let the categorisation into kinds be ‘can exist in the population’ and ‘cannot exist in the population’ as applied to points in between these extremes. By the continuity hypothesis, small changes in the value of the parameter ‘how homosexual is an individual’ do not change the property of existence, so we can apply the strong sorites argument to fill in the gaps. So I conclude that sexual orientation is a continuum.

Classifying people into kinds as homosexual or heterosexual and using the continuity hypothesis again gives rise to a sorites problem, and here the classification is not obviously always valid (which we know, because of the debatable middle region), so we end up with three regions: predominantly-homosexualpredominantly-heterosexual and mixed sexuality.

Sex, gender and sexuality are independent

I want to show that the three attributes sex, gender and sexuality are independent. To simplify the discussion I will start by showing that this is true of all three pairs of attributes, and then use this to deduce independence for the triad.

Sex and gender are independent

It follows from Jungian gender theory that gender is largely independent of sex, and may even change over time, as the individuation process proceeds. In many cases the subconscious gender component will be undeveloped, either through lack of self-awareness, or due to societal pressures to conform to stereotyped sex-specific behaviours. This is undoubtedly the reason why it took until the twentieth century for it to be fully understood that the gender that one thinks one is is not necessarily the same as ones sex. As examples consider people who decide to undergo gender reassignment (ignoring, as I said earlier, the small number of cases of people whose biological sex is confused): I would argue that these are individuals whose unconscious gender component (so the anima in a man) is so strongly developed that they simply cannot identify with their bodies, and so feel compelled to change their sex to the one that fits their self-identification.

Sex and sexuality are independent

That sexuality and sex are independent is trivial: homosexual preferences are not the prerogative of one sex, and neither are heterosexual preferences.

Gender and sexuality are independent

The independence of gender and sexuality, while non-trivial, is also relatively simple to prove from direct observation (so there is no need to rely on analytical psychology). At the homosexual end of the spectrum there are the well-known butch and femme gender roles. At the heterosexual end there are both masculine and feminine individuals. If we imagine gender and sexuality as being set out as the axes of a square, this gives us the four corners.

I want to apply the strong sorites argument with the categorisation into kinds ‘can exist in the population’ and ‘cannot exist in the population’. Using the continuity hypothesis, I can apply the strong sorites argument, so all combinations are possible.

Note, as a final critical point that the basis of this argument, i.e. the existence of male-ish and female-ish individuals near both ends of the sexuality spectrum is independent of sex. There are butch and femme lesbians and gay men, and there are masculine and feminine heterosexuals of either sex. Thus the whole gender / sexuality spectrum exists independently for each sex.

Sex, gender and sexuality and independent

Pick a sex. For that sex we have shown that all possible genders and sexualities can exist. Moreover, we have shown that all possible combinations of gender and sexuality can occur in that sex. Repeat the argument for the other sex. Independence is now demonstrated.

Analysis 2: sexual & aesthetic preferences

Finally we reach the core of the essay: the discussion of how sex, gender and sexuality effect aesthetic preferences in terms of to bodies of which sex(es) is an individual attracted aesthetically. The argument proceeds in two stages. First I tackle the (relatively) tractable problem of determining how these three variables influencesexual preferences, i.e. to bodies of which sex(es) is the individual attracted sexually. This done, second I turn to the much more complex question of aesthetic attraction; complex because (as I will show below), though many theorists (including those who formulated the original ‘male gaze’ theory) treat sexual and aesthetic attraction as if they are identical, they are not. They are very similar, but it is possible to have sexual feelings for a person one finds unattractive, and to find attractive a person towards whom one has no sexual feelings.

Throughout this discussion I deal with what one might call instinctive preferences, that is to say those determined by the structure of an individual’s psyche. There are, however, also conscious preferences: an individual may decide, consciously, to find a particular body beautiful despite not being naturally drawn to it (this, after all, is how fashion works), or that, for entirely rational reasons, they should make another individual their sex-partner (e.g. arranged marriages). As conscious preferences involve intellectual decisions, I cannot speak of them within my framework, which works in terms of the unconscious mind, and so, bearing Wittgenstein’s dictum in mind, of them I shall remain silent. Therefore, note that all of the following discussion relates only to instinctive preferences.

Sexual attraction

Which factors are relevant to sexual attraction?

Sexual attraction is an expression of desire to have sexual relations with an individual, and desire is a psychological state. Thus the sexes of the individuals to whom I am attracted will be determined by the structure of my psyche. As such it can depend directly on psychological variables, but only indirectly on physiological factors like sex. By which I mean that physiological factors may contribute to psychological variables (e.g. via the action of hormones on the brain), but these psychological variables will not be determined solely by physiological factors. As we see with even the most basic of physiologically driven psychological variables – pain – other psychological factors can modify and even over-ride the physiological factor. Thus rather than working with (potentially) relevant physiological factors, we should identify the relevant psychological variables. Therefore, provided we understand which relevant psychological variables it influences, we can remove sex from the equation (for the moment).

Start with the two variables discussed at length above. Obviously gender is influenced by sex: as an individual’s gender is how they themselves identify their sex, sex is bound to be a contributing factor, even though other factors can, and do, over-ride it. As for sexuality, there is some evidence that homosexuality is slightly more prevalent among men than women, but given that we have shown that any attempt to categorise people as either homosexual or heterosexual is doomed to fail, it is not clear how much credence would should place in this. However, using the principle of charity, let us accept that there is a weak dependency of sexuality on sex.

We leave for future study the question of whether any other psychological variables impact on sexual preference. A model based on gender and sexuality is sufficient for our purposes.

How is sexual attraction determined?

The naive answer to this is that sex plus sexuality determine the sex(es) to which an individual is attracted. This view is inherent in the ‘male gaze’ theory’s assertion that art disproportionately represents female bodies to cater to the desires of heterosexual men. However, we have seen that sexual desire is a psychological construct. Psychologically, an individual’s ‘sex’ is determined not by their actual physiological sex, but by the sex that they conceive of themselves as being, that is their gender. Thus it is surely more correct to say thatgender plus sexuality determine the sex(es) to which an individual is attracted.

Let us explore some consequences of this by looking at extreme cases (reintroducing sex as a factor to make the consequences of the hypothesis more concrete):

  1. A purely masculine heterosexual man.  This individual will be sexually attracted exclusively to women, and so forms the traditional ‘male’ stereotype, as per the ‘male gaze’.
  2. A purely feminine heterosexual man. This is more interesting, as the individual will be attracted exclusively to men despite being heterosexual. Two outcomes are possible: the individual is a transexual and reassigned as a femme straight woman, or the individual becomes a femme gay man.
  3. A purely masculine homosexual man. This individual will be sexually attracted exclusively to men, and so will be a stereotypical butch gay man.
  4. A purely feminine homosexual man. This individual will be sexually attracted exclusively to women, and so will appear, to all intents and purposes to be a femme straight man. However, as the conflict of body with gender will most likely lead to psychological unease, a more likely outcome is transgendering, with reassignment as a femme lesbian woman.

Examples are all for a man purely for reference (not because of any preference); for a woman interchange genders mutatis mutandis. It can be seen that as we look at both sexes, all eight combinations of butch / femme + straight / gay + man / woman occur. In particular, the treatment of transgender is very comprehensive.

This is rather interesting, as it implies that apparent sexuality need not always be the same as intrinsic sexuality: apparent sexuality is the result of a combination of intrinsic sexuality with gender. This is because sexuality is about attraction to bodies, i.e. sexes, whereas the thing that determines which bodies that sexually attracts one to is not one’s sex but one’s psychological sex, i.e. gender.

To reiterate, the reason this seems counter-intuitive, or even just plain wrong, is that we tend to assume that bodies attract bodies. But in fact, bodies attract minds, and minds need not necessarily consider themselves to be of the same ‘sex’ as the body they inhabit. Once this is accepted, the surprising, but, on a second look rather compelling, outcomes described above arise. And the more convincing seems my model:

preferred sexsexual orientation applied to gender

Classifying types of preference

A quantitative analysis of the influence of gender and sexuality on preference arrives at the following result. If an individual is masculine with probability g and homosexual with probability s, then the probability of their being attracted to the two sexes is:

Attracted to men with probability = 1 + 2gs – g – s

Attracted to women with probability = g + s – 2gs

The derivation of these results can be seen easily from the following figure:

Interpreting the mathematics, we deduce the following:

  • At the extremes of the ranges (so g and s are both equal to 0 or 1) we get ‘normal’, i.e. exclusive, homosexual or heterosexual behaviour.

  • Along the four edges (i.e. where one of the two variables is equal to 0 or 1) the formulae give probability of attraction equal to the other variable or 1 minus the other variable (this is to be expected, given our analysis or exclusive cases above, where we showed that flipping either sexuality or gender reversed preferences). So on the edges, preference is defined by only one variable.

  • If g is equal to one-half, then in both formulae the resulting probability is one-half, irrespective of sexuality. Similarly, ifs is one-half, then in both formulae the resulting probability is one-half, independent of gender. So if either gender or sexuality is (close to) evenly balanced, the individual’s preferences will be entirely unpredictable.

In the four ‘in-between’ regions the probabilities interpolate between these values, the interpolation taking the form of parabolic curves. This can all be visualised as follows (the graph shows the probability of attraction to women):

Thus we conclude that there are nine regions in total:

  • Four regions close to the corners, where both gender and sexuality are exclusive, so preference is exclusive. These can be labelled predominantly-man-preferring and predominantly-woman-preferring regions.

  • A region where one or both of gender and sexuality is close to being equally balanced between the two extremes. This is the region of mixed-preference, where we can say nothing at all about the preference of individuals falling within it.

  • Four regions of approximate-preference, where the preference is predominantly for one sex, but with an admixture of the other. They can be described as falling into two classes: approximately-man-preferring and approximately-woman-preferring.

So there is a five-fold typology:

  1. Predominantly-man-preferring
  2. Approximately-man-preferring
  3. Predominantly-woman-preferring
  4. Approximately-woman-preferring
  5. Mixed-preference

This is illustrated as follows:

So what do we do with all this? There seem to be three approaches:

  • Keep full generality and use the underlying quantitative model given by the probabilities. This is very useful for detailed studies, but not for the relatively coarse-grained work we are interested in: there is little point in defining a ‘43% male, 37% homosexual gaze’!

  • Adopt a five-fold typology: predominantly-man-preferring, predominantly-woman-preferring, approximately-man-preferring, approximately-woman-preferring and mixed-preference. This complicates things somewhat, but it has the merit of being the best qualitative representation of the underlying quantitativemodel.

  • Merge approximately-man-preferring with predominantly-man-preferring, merge approximately-woman-preferring with predominantly-woman-preferring, thus obtaining a three-fold typology: predominantly-man-preferringpredominantly-woman-preferringmixed-preference. This trades off accuracy against simplicity.

We adopt the third approach, as it is all that we need for gendered gaze theory. However, it is recommended that subsequent work investigate the full five-fold typology (or, better still, the underlying quantitative model).

Aesthetic attraction

As soon as we try to switch from looking at sexual attraction to aesthetic attraction problems arise. Naively one might believe the two to be strongly related, but that isn’t the case. Yes, there is a relationship, but a little thought can show that it is possible to be sexually attracted to individuals who one does not consider beautiful (the sexiest woman I have known was not remotely beautiful, and yet the sexual allure she generated was enormous) and to consider beautiful individuals who we do not consider sexually attractive is equally possible (I find Virginia Woolf beautiful but absolutely un-sexy). So what can we do?

This problem is symptomatic of the point I made above distinguishing instinctive and conscious preferences. Sexual preference is overwhelmingly instinctive: we know when we are attracted to an individual because our bodies respond without our appearing to tell them to do anything of the sort. Aesthetics, almost by their very nature involves conscious preferences. Aesthetic values are undoubtedly, at root, based on sexual preference (so if one could find a naive individual, who had had no cultural contact with other individuals they would desire those they found attractive and vice versa), but overlaid on that are the very powerful input of culture and deliberately learned taste. We even speak of things as ‘acquired tastes’, meaning that they are learned. So, for example, I find Francis Bacon’s horrifyingly distorted figures beautiful, because I have cultivated a taste that allows me to do so. Overall, this is perhaps not surprising in view of the much-commented-on fact that the current ‘desirable’ shape for women in western culture seems to be completely at odds with the kind of womanly shape that is sexually desirable.

What this means is that there is no way that the theory developed above can be extended from sexual to aesthetic preferences, as any model for the latter involves not just modelling the (relatively) simple unconscious mind, but potentially all of the conscious mind. The best we can do is to say that there is some (strength unknown) correlation between sexual and aesthetic preference. Therefore gendered gaze theory can be stated as follows:

Gendered gaze theory

Individuals can be categorised based on their sexual attraction to other individuals as follows:


  • Predominantly-man-preferring

  • Predominantly-woman-preferring

  • Mixed-preference

Which class an individual falls into depends on their gender and sexuality in a relatively complex, but understood, way. There is a statistical correlation between the sex of figures appearing in art-works created by an individual and which of these classes the individual falls into. The strength of the correlation will vary from individual to individual.

It is hardly surprising to find that this is a much more circumspect conclusion than that made by the theorists of the ‘male gaze’. This may seem like a defeat, but is cannot be entirely negative to replace an over-confident assertion with one that is more nuanced.


One response to “The Gendered Gaze

  1. Amro Gebreel 18/02/2011 at 11:19

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: