Mechanisms and Deficits of Face Recognition
2500 years ago the ancient Greek philosopher Plato developed his theory of Ideas. He seemed to suggest that there is a form of beauty, ideal beauty, based on which we judge other things as either being beautiful or not. To put this in more simplistic terms he suggested that a chair is a chair and it is not a dog because there is an ideal, perfect, form of chair by which we judge chairs in general and there too is an ideal, perfect form of a dog by which we judge dogs in general. He went so far as to suggest that these perfect forms exist out there somewhere detached from ourselves.
2500 years later science confirmed the theory of Plato, but instead of calling it forms they called it schemas. Maybe science has been unable to confirm whether these forms 'exist out there'. Or to put it in modern terms, to confirm whether we are born with an innate ability to distinguish between beauty and ugliness, or whether we learn what beauty is. It has been mentioned that a child tends to stare longer at beautiful faces, maybe this would be enough to confirm the second theory of Plato.
Plato did not go so far as to philosophise on how we differentiate people or chairs from one another. On our everyday life we treat chairs from people very differently. A chair merely serves a purpose and we might judge it on whether it is aesthetical or comfortable. We might scrutinise it for the first five minutes and maybe praise it for being so nice, but then we just move on and the chair is just another chair.
When it comes to people, we scrutinise them every second we are with them and never stop doing so. While looking at a chair we might be in awe for five minutes, looking at a person gives us a variety of emotions. While we might see a chair upside down, we hardly ever get to see a person upside down. While most of the time we just see a chair and sit on it, we see a person and within seconds analyse all of his features and even the impression that those features give.
With all that being said, is an object really different from a face? That is what we will be looking at in this essay. We will start by discussing whether faces are the 'special' class of objects, whether we process them in a configural or componential way, the modules that are involved in face recognition and conclude by discussing what makes a face attractive.
We will start this essay with describing the model of Bruce and Young (1986) of face recognition. Unfortunately it seems impossible to have access to the original article of Bruce and Young: 'Understanding Face Recognition'. According to this model face recognition starts with the visual encoding of the features of a face, meaning that one is able to identify a face viewed from any angle through its features (Hook 2004). The model then differentiates between facial familiarity and facial expression. This is supported by cases of brain injuries whereby a person is able to identify muscular movements on a face but unable to attach any meaning to them. In other words, the person is able to recognise faces, but not facial expressions (Young et al. (1993), on the other hand Tranel et al. (1988) has reported patients with the opposite ability. The third stage is the most important as in this stage, once the features of a face are analysed, the person then stores them into what Bruce and Young called 'face recognition units' which are then connected to the 'person identity nodes' which as the name suggests contains personal information about this person (Snowden, Thompson and Neary 2004). While this model might give us the neurological basis of how face recognition occurs and while it might tell us about the processes involved in recognising a face, it does not tell us much about what differentiates a face from another. It mentions that a person stores the features of a face, but it is silent in showing how these features interact with each other and how exactly are these features analysed.
To answer these concerns we need to look at configural and componential processing. The Thatcher illusion seems to support the configural processing of faces in the famous experiment whereby a person can not see the grotesque modifications of Mrs. Thatcher's face when the picture is inverted, while when the picture is upright the modifications are apparent. Other experiments seem to support the configural processing of faces over that of componential processing. For example, Leder and Carbon (2004) state about the results of their experiment that: 'When full faces had been learned, the expected finding of advantageous recognition of holistic presentations was replicated (Exp. 1). However, when only facial parts had been learned (Exp. 2), this effect was reduced and even reversed, indicating that holistic superiority with both sorts of faces depends on holistic learning strategies.'
The inverted effect on facial recognition might be seen as evidence that faces are 'special' classes of objects and different from other objects. It has however been suggested that this may not be entirely correct. People are expert in facial recognition as we see faces every day countless of times and scrutinise them every instant. This can not be true of other object or animals as we might not spend too much time or as much time scrutinising them as we do faces. To conclude whether faces are a 'special' class we can not consider inverted faces with inverted objects to be equivalent. On the other hand, considering inverted faces with pictures of inverted dogs presented to experts who have for years worked with dogs or judged them in dog shows we might have an equivalent comparisons. Such experiments have been carried out and Diamond and Carey (1986) concluded that: 'No, faces are not special.'
Support however for the 'specialty' of faces as different from objects has been found from cases of patients who are unable to recognize faces, but they can recognize objects. Prosopagnosia has been studied and tested in countless experiments and the results have been taken to support the view that faces are different from objects. Nonetheless Gauthier, Behrmann and Tarr (1999) argue that the effects of prosopagnosia are due to 'subordinate-level face recognition'. As mentioned above people are experts at recognizing faces, hence the difficulty of prosopagnostics in recognizing faces might have more to do with losing this expert ability rather than faces being distinctively different from other objects. It has been shown that the 'face area' which affects the prosopagnostics is activated in 'subordinate-level categorization of nonface objects'. Another finding to support this view is that in the experiment by Gauthier et al. (1999) the prosopagnostic participants were 'as sensitive with faces as with Greebles' which led them to suggest that maybe 'rosopagnosic subjects could have lost the ability to use such expertise', the expertise that normal participants have when judging faces.
It was mentioned above that Bruce and Young (1986) suggest that facial expression is independent of facial recognition and they run in parallel. To support this claim cases of patients who are unable to recognize faces but can recognize expressions or the opposite can be mentioned. Other everyday examples can be used as well as 'we sometimes have to look twice to recognize someone familiar if she displays a facial expression never seen before or we may get confused when a seemingly strange person smiles at us.' (Wild-Wall, Dimigen and Sommer 2007) Faces generally tend to activate the 'fusiform gyrus', otherwise called the 'face area', which tends to be activated when processing faces. Facial expression processing has been shown by Haxby et al. (2000) to include 'occipitotemporal cortex with projections to the fusiform gyrus and superior temporal sulcus.' Differences of 'event-related potentials' for facial recognition and expression has been shown by Bobes et al. (2000) which has been taken to support the view that facial recognition is different from interpreting facial expressions.
Why would facial expression activate different areas from facial recognition? One reason might be that facial expressions are emotionally charged which might not be true for recognising faces which has been supported by the fact that facial expression has been shown to activate the amygdale (Sato et al., 2001).
Faces can be used to express emotions as well as other characteristic. The most intriguing of which might be what makes a face attractive. In an experiment by Rhodes, Geddes, Jeffery, Dziurawiec and Clark (2002) it was found that children are not born with an innate sense of beauty. The experiment shows that they are more inclined towards novel looking faces rather than 'attractive' faces. This is supported by the fact that a child tends to look longer at less average faces and the asymmetrical faces which is in contrast with adults who tend to look more at average and symmetrical faces. There is much to be desired with this experiment however, as attractiveness is derived from 'Data from the mothers'. Are people who are not familiar with the objectivity required, how to eliminate confounding variables etc. to be trusted? Other experiments have shown that infants generally tend to look 'more at attractive faces when these are shown paired with faces judged by adults to be unattractive' (Slater et al. 2000).
Whether we are born with an innate ability to perceive beauty, as Plato suggests, seems to be the case. Nonetheless one must be weary to conclude so as very little is understood about it. One thing which seems to be understood is the fact that adults generally tend to like attractive faces and symmetrical faces. This might be so due to our evolution as unattractiveness might be connected with unhealthy genes, or maybe those who did not prefer average faces might have struggled to find someone superior in terms of beauty hence failed to reproduce. It also could be said that symmetry is another expression of healthy genes which might explain why adults prefer symmetrical faces.
