Background Considerations To Facial Esthetics
Background Considerations To Facial Esthetics
Background Considerations To Facial Esthetics
Abstract. This paper examines the nature of our objectives in attempting to improve facial appearance. Questions are
addressed concerning the basis for any collective agreement on ‘improvement’ and the nature of an agreed ‘ideal’, includ-
ing an exploration of the reasons for its existence. The discussion focuses on the concept of ‘averageness’, as well as
supplementary hypotheses. Finally, the origins and validity of contemporary clinical guidelines are addressed.
Index words: Aesthetic standards, Averageness, Multiple-motive hypothesis, Secondary sex characteristics.
the process preferring mainly ‘non-dominant’ features. that the most attractive faces additionally have these extra
Similarly, Cunningham’s (1986) experiments on the features, i.e. those shown in Figure 2. They argue that other
relationship between specific female features and male factors beyond averageness are in operation since:
response indicated that men were especially attracted to
1. In the natural environment, ‘non-average’ male animals
faces demonstrating those features that Keating had
are often the most successful at mating and producing
demonstrated as ‘non-dominant’ (see Figure 2) or ‘neo-
offspring, i.e. those with greatly enhanced secondary
natal’ in type, having large foreheads, large, wide set eyes,
sex characteristics (for example, ornamental such as
small nose and chin, and full lips. It was suggested that these
brightly coloured plumage, e.g. peacock feathers, or as
characteristics tended to stimulate a nurturing or caring
weapons, e.g. deer antlers). Similarly, women might be
instinct in the male observer as these features are particu-
expected to prefer males with evidence of increased
larly characteristic of infants (Enlow, 1990). However, the
muscularity or robustness rather than those with merely
most attractive female faces of all, had additionally some
average characteristics.
‘maturity’ features such as prominent cheekbones. In
2. Langlois and Roggman’s (1990) data did show that
Cunningham’s words, ‘. . . the contribution of mature with
there were a few exceptionally attractive individuals,
neonate features may signal that the female is at an optimal
who were more attractive than the composites.
age for mating . . .’.
3. The process of producing composites from a large
In their study on female perception of male facial
number of individuals inevitably eliminates facial skin
attractiveness, Cunningham et al. (1990) concluded that
blemishes including creases, wrinkles, etc., thus pro-
females were attracted to males who had dominant or
viding the composite face with an ‘unfairly’ clear com-
mature characteristics such as wide jaws, strong chins, and
plexion.
relatively thin lips, but that the most attractive men
additionally had some neonatal features (as in Figure 2), i.e. Additionally, there is some evidence that attractive com-
expressing both ‘ruggedness’ and ‘cuteness’. Additionally, a posites can be made even more attractive by further
third feature, namely ‘expressiveness’ was also present manipulation (Perrett et al., 1994).Accordingly, other factors
amongst those male and female faces judged to be the most such as enhanced facial secondary sex characteristics appear
attractive, and these would include width of smile, and the to be involved, especially in males, where these will include
presence of high-arched eyebrows, indicating sociability such ‘dominant’ features as a large jaw, strong chin, etc. So,
and responsiveness. Thus, these three features—neoteny, if the ideal that clinicians are striving to achieve (albeit
maturity, and expressiveness—all contribute to the ‘multiple- unknowingly) is a sexually modified form of ‘averageness‘,
motive’ hypothesis of attractiveness, indicating that male how does this come about? In other words, why should this
attractiveness is based on a combination of these three be the common denominator for facial aesthetics?
features with maturity or dominance particularly empha-
sized, whilst female attractiveness places more emphasis on
neoteny.
Sexual reproduction and parasite resistance
This theory and that of ‘averageness’ are not necessarily
mutually exclusive. Alley and Cunningham’s (199l) com- Needless to say, the reason for a modified form of
mentary points out that average faces are attractive, but ‘averageness’ being the aesthetic ideal is not to be found in
the orthodontic literature, or indeed any of the surgically females, indeed those individual males that rated more
related texts. Instead, one has to examine the zoological highly for attractiveness than the composites, had these
literature, where the assumption is clearly made (e.g. very characteristics.
Thornhill and Gangestad, 1993) that human perception of It is also clear that symmetry is an important indicator of
attractiveness is firmly based in evolutionary biology, in freedom from disease, providing evidence that the indi-
that it is an essentially sexual characteristic, based on the vidual is in good condition and thus a worthy potential
need to find the most suitable mate for the production of mate. Needless to say, the reverse is equally true, so that
healthy offspring. In the natural environment, parasites are what is termed ‘fluctuating asymmetry’ (indicating its
constantly evolving in a bid to break down their potential haphazard nature, and in distinction to some forms of
hosts’ defences and the hosts, in turn, are also evolving, deliberate natural asymmetry) is a biological signal that the
albeit at a slower rate, to improve their defences in a form phenotype has been subjected to stress in some way and so
of natural ‘arms-race’.The strongest defence that a host has is unlikely to provide his mate with healthy offspring. For
is its genotype, so that a heterozygous host will have a better example, environmental stress in the form of fowl mite
genetic code or ‘good genes’ (Hamilton and Zuk, 1982) concentrations were shown by Moller (1992) to produce
than a host derived from in-breeding. Usually, the female is asymmetry in the length of the tail feathers of male barn
the one who actively selects the mate (males being rela- swallows. For this particular bird, the tail feathers are
tively indiscriminate) and the female will often go to some considered to be the secondary sex characteristics and the
lengths to ensure that her potential mate shows evidence of production of asymmetry in their length would impair its
pathogen or parasite resistance. Indeed, Hamilton and Zuk chances of finding a suitable mate. Parsons (1992) has
(1982) likened this phenomenom to that of a family doctor described how fluctuating asymmetry can be used as a
examining a patient for eligibility for life insurance and that means of monitoring the degree of stress undergone by a
careful examination of, for example, plumage or fur quality particular phenotype, stress being applied environmentally
may be undertaken and in some cases, urine or faeces may (as shown by Moller in using parasites), or equally through
be inspected. Hamilton and Zuk further showed that there extremes of temperature, protein deprivation, or noise,
was a very clear, inverse relationship between parasitaemia combinations of such stressess provoking asymmetries in
levels in varieties of North American birds with brightness mandibular and maxillary first molar sizes in Wistar rats
of plumage—particularly in males, where plumage colour is (Sciulli et al., 1979).A different type of stress that can result
a secondary sex characteristic. In other words, the presence in asymmetry is genomic, i.e. associated with ‘in-breeding’,
of brightly coloured plumage on a male bird would indicate producing homozygosity. This has been shown to produce a
to a female that the male is relatively free of parasites and variety of asymmetries (Clarke et al., 1986; Clarke and
so is likely to prove to be a suitable mate for offspring MacKenzie, 1987). The implication is that the somewhat
production. Other qualities, such as size and vigour, simi- asymmetric phenotype is perceived by a potential mate as
larly prove useful indicators of suitability in terms of unsatisfactory. Needless to say, symmetry and averageness
warding off competitors, provision of appropriate breeding are closely related factors in the perception of attractive-
facilities, etc. ness and to biologists such as Thornhill and Gangestad
This hypothesis, namely that secondary sex character- (1993), human perceptions have a clear evolutionary bio-
istics advertise parasite resistance and thus the presence of logical basis.
‘good genes’ has been the subject of many zoological In summary, facial attractiveness is essentially a prelude
investigations since the publication of Hamilton and Zuk’s to sexual selection and thus to reproduction. In other
paper. The introduction by Hausfater and Thornhill (1990) words, its biological basis is as a means of providing off-
and, indeed, all the subsequent papers in that special spring. The basis for sexual selection is health, i.e. parasite
edition of American Zoologist, were devoted to this topic, resistance, which is ‘advertised’ by enhanced secondary sex
the conclusions conferring a considerable amount of valid- characteristics, in the human male by those aspects of facial
ity to the concept. Following this, Folstad and Karter (1992) growth enhanced by testosterone at puberty, i.e. jaw size,
presented a further explanation for the role of testosterone chin prominence (within limits), and to some extent, malar
that, of course, is responsible for the production of second- eminence, whilst in the female this is limited to malar
ary sex characteristics. In introducing their ‘immuno- prominence alone. As mentioned earlier, parasite resist-
competence handicap’ hypothesis, they pointed out that ance is enhanced by heterozygosity, which in turn is
testosterone has a known side-effect, namely that it reduces ‘advertised’ by averageness—particularly, in humans, in the
the male animal’s immunocompetence, potentially render- female. Thus, sexual selection favours both averageness
ing it more susceptible to pathogens. Only a ‘strong’ male (especially in the human female) and enhanced secondary
could therefore cope with this. Thus, by a process of ‘honest sex characteristics (especially in the human male).
advertising‘, a male animal with highly developed second- It would appear that there is an evolutionary basis for our
ary sex characteristics is essentially announcing to potential perception of attractiveness and thus for our objectives as
female mates that his genetic make-up is so strong, that he clinicians. Accordingly, our objective is a symmetrical face,
can afford the mild immunocompetence handicap that the with appropriate enhancement of secondary sex charac-
production of his testosterone is producing and still be teristics and, most importantly, with dimensions very close
healthy. Grammer and Thornhill (1994) indicate that in the to the mean of the population. With this in mind, it is
human male, those facial characteristics such as a particu- appropriate to consider the merits of the guidelines or
larly wide mandible and strong chin are, in fact, the test- recommendations for this ‘ideal’, which one comes across in
osterone-produced human equivalent of secondary sex the literature.
characteristics. This would help explain why these features
when present (within limits) tend to rate highly amongst
JO June 2001 Features Section Background Considerations to Facial Aesthetics 163
Aesthetic standards—sources of evidence earlier Greek sculptors had been very much aware of
‘averageness’ in aesthetics, pointing out: ‘. . . whatever form
Guidelines for the ‘ideal’ in facial attractiveness have been
is most beautiful in man or in the cow or in the lion, always
presented in a number of ways, namely: artistic, cephalo-
looking for the mean within each Genus’ (Olds, 1993, refer-
metric (including growth study material), and anthropo-
ring to Pollitt, 1965). A famous sculpture by Polykleitos
metric.
called ‘The Lance Bearer’ (Figure 3a) conforms to certain
carefully laid down proportions, for example, the head and
Artistic guidelines face together comprises two of fifteen equal vertical seg-
ments into which the human form could be divided, a
It was only as recently as the mid-1970s that ‘true’ evidence principle which holds true today. The head of Aphrodite
in the form of mean data derived from growth studies (Figure 3b) symbolizes the ideal facial proportions as
became available for clinical use. Until then, those inter- envisaged by Greek sculptors in the fourth century BC.The
ested in improving an individual’s facial appearance had to concept of the so-called ‘golden’ or ‘divine’ proportion
rely on guidelines based on the experience of represent- (Ricketts, 1982) also developed in the same period. This
ational artists over the ages, notably the Renaissance artists concept is believed to have orginated with the sculptor
during the fifteenth and sixteenth centuries. Whilst the Phidias, hence the expression ‘phi’ in relation to the golden
ancient Egyptians were possibly amongst the first to try to proportion which has been used as an aesthetically pleasing
describe facial and bodily proportions in mathematical or relationship of vertical and/or horizontal structures, from
in grid form, the main contribution seems to have developed the Parthenon to contemporary design (viz credit cards).
during the ancient Greek civilization, particularly during More to the point, as indicated by Ricketts, this relationship
the fourth and fifth centuries BC. Sculptures of the human seems to occur naturally in a variety of guises in the human
form produced during that period invariably showed pro- face and body. Whilst the Romans relied on the work of the
portions that conform to certain established rules or Greeks, there are some significant original Roman contrib-
‘canons’. Guidelines were laid down by a number of artists, utors, such as the architect Vitruvius. His famous facial
one of the most famous, Polykleitos in the fifth century BC, trisection (Figure 4) is still recommended in modern texts
whose statue and accompanied written work (Pollitt, 1965) (e.g. Powell and Humphries, 1984), and used by ortho-
displayed a number of recommended ratios for the human dontists and orthognathic surgeons today, i.e. 2000 years
figure. Whilst the ‘canon’ itself no longer exists, many of later. During the thousand years that followed the Roman
Polykleitos’ recommendations remained and were later Empire, artists moved away from depictions of ‘ideal’
modified and copied by the Romans. The second century human form as the concept of physical beauty gradually
Roman physician and philospher, Galen, noted that the became unacceptable as an ‘ideal’ in an increasingly
F I G . 3 (a) Polykleitos’ ‘The Lance Bearer’, fifth century BC. The head and face comprise two of the 15 equal vertical parts dividing the figure. Reprinted by
permission of Charles C Thomas, Illinois. (b,c) Aphrodite, fourth century BC. Reprinted by courtesy of the Museum of Fine Arts, Boston.
164 R. J. Edler Features Section JO Vol 28 No. 2
(a)
F I G . 5 (a) Drawing of the proportions of the head (and eye). Da Vinci F I G . 6 Sagittal and vertical relationships within the face:
(date unknown) reproduced by permission, Bridgeman Art library. (b) The face (a) Da Vinci – proportions of head, C1488–90, R L 12601, The Royal Collection
divided equally into vertical fifths, Powell and Humphries (1984), reproduced c. 2000, Her Majesty, Queen Elizabeth II, Royal Library, Windsor Castle.
by permission. (b) Durer, reproduced by permission, J. Pierpoint Morgan library.
JO June 2001 Features Section Background Considerations to Facial Aesthetics 165
These and other criteria are important in assessing validity Anthropometric evidence
and were only really fulfilled this century, following the
introduction of cephalometry, and subsequent growth Direct measurements taken from living subjects would
studies and by anthropometry. seem an obvious source of valuable data. As pointed out by
Ward (1989) and Rogers (1974), the advantages of this
approach include its non-invasiveness, simplicity, cheapness
and, above all, the fact that there is a comprehensive dataset
of age- and sex-matched standards (albeit presently limited
Cephalometry to Caucasians) to which patients’ measurements can be
A discussion on the background to facial attractiveness compared. In assessing the validity of direct measurements
must obviously acknowledge the enormous impact that taken with sliding and spreading calipers and tape, Ward
cephalometry has had and the manner in which it has been and Jamison (199l) found the magnitude of error amongst
used to provide a vast array of data useful for the rep- most of the 52 craniofacial variables studied to be low, i.e.
resentation of ideal proportions. However, to those inter- within a millimetre. Significant errors related inversely to
ested in finding a dataset that could accurately describe the size of the measurement and, just as in cephalometrics,
an ideal model for facial attractiveness, there are three landmark identification, so that smallish measurements
areas of deficiency in cephalometrics that need to be associated with landmarks that would be difficult to
addressed. identify, i.e nasal root width, are admittedly, not particularly
First, in many cephalometric analyses, the population reliable.
details from which mean data was derived is not always Koury and Epker (1992) have summarized the key
clear. Where the size of the sample is known, there can be anthropometric measurements of value to our clinical area,
enormous variation, e.g. Ricketts analysis which involved using measurements derived from the extensive work of
the examination of 1000 individuals (Ricketts, 1957) to Farkas (198l) and Farkas and Munro (1987).The population
Downs analysis, involving just 25 (Downs, 1948). In some sample of Farkas totals over 2500 Canadian subjects, the
instances, for example, the analyses of Holdaway, Jarabak, measurements having been taken during the period
and Steiner, the population size is unknown and this intro- 1967–1984. The subjects were all Caucasian, a mixture of
duces an element of doubt when considering the validity of Anglo-Saxons, Latin, and those of Germanic origin (Farkas,
suggested guidelines. Despite this, one has to acknowledge 1987). Thus, currently the value of anthropometric
that some of the guidelines, for example Steiner’s aesthetic measurements is limited to Caucasians, but for this group,
plane (possibly based on just one patient), have undoubt- there is enough evidence of mean measurements, both
edly proved to be extremely useful. linear and also as ratios or proportion indices, with which to
Secondly, most of the data presented are two-dimen- compare one’s patients. This evidence has been used in an
sional in nature and thus incomplete. Measurements to examination of the validity of previous artistic recom-
represent the transverse dimension, i.e. as taken from PA mendations such as the work of the Greek and Renaissance
cephalometric radiographs are, of course, entirely dento- artists (Farkas et al., 1984, 1985a,b). For example, the ‘true’,
skeletal and yet it is the soft tissue appearance that the i.e. anthropometric relationship of upper/lower lip lengths
patient presents to the outside world. In other words, at in relation to lower face height is compared to the recom-
the present time there are no comprehensive, cephalo- mendations of two Renaissance artists, as illustrated in
metrically-based three-dimensional soft tissue measure- Figure 7 (Farkas et al., 1984). Figure 8 illustrates the com-
ments available to represent the mean, nor as yet from posite ‘ideal’ faces constructed from neoclassical ‘canons’,
any other imaging techniques such as video or optical i.e. ‘Renaissance Man’ and from Farkas’ mean data. In this,
scanning. one can see that, for example, the Vitruvian trisection is
Finally, whilst acknowledging the value of cephalometric perhaps not absolutely correct, at least in the male whose
growth study material such as that of the Burlington lower face height is a little greater than his mid-face height,
(Popovich and Thompson, 1977), Michigan (Riolo et al., which in turn is greater than his upper face height. Addi-
1974) and the Cleveland (Broadbent et al., 1975) studies, tionally, his nose tends to be shorter than his ear and the
Proffit’s (2000) comment is relevant, namely that these interocular distance is greater than the eye fissure length, in
studies were undertaken as long ago as the period 1935– contradiction to previously held canons. Of course, the
1965, raising the question as to their current validity, significance of the difference in ethnicity between the two
bearing in mind population changes that have since occur- populations, separated by 500 years, has to be acknow-
red. In other words, the most well known growth study ledged.
material is, as Proffit describes, historical in nature. Bhatia
and Leighton’s (1993) growth study material is more up-to-
date, their study having begun in 1952, but despite its great
value in providing norms, especially those for orthognathic
planning, it is still two-dimensional in nature.
In order to provide relevant information to facilitate
aesthetic improvement, what seems to be needed is a
comprehensive, up to date dataset of soft tissue-based
F I G . 7 Relative proportions of upper lip, lower lip, and chin lengths.
measurements, which need to be three-dimensional in Anthropometric proportions (Farkas et al., 1984) compared to the guidelines of
nature. Interestingly, all these criteria are more than Renaissance artists. Anthropometric proportions of the upper lip agree with
adequately fulfilled by a source of evidence that has been those of Francesca and Pacioli, whilst lower lip lengths coincide with Durer’s
neglected for perhaps too long, namely, anthropometry. guidelines. Reprinted with permission, American Journal of Orthodontics.
166 R. J. Edler Features Section JO Vol 28 No. 2