Reflecting on mirror-writing

The article is the most read piece on our website ever; viewed hundreds of thousands of times, and still getting around 10,000 views a month. We caught up recently with the authors, to talk about the article's success, and to get an update on any developments in the field.

Let's start with a definition. What is mirror-writing?

Reversing letters, numbers, words or entire phrases when writing is called mirror-writing, because these letters would appear normal when viewed in a mirror. Mirror-writing arises in three main ways. It is common, even ubiquitous, amongst children learning to write. It can appear abruptly in adults, often after damage to the left side of the brain, or at times of extreme psychological stress. And sometimes, it is done deliberately. The most famous practitioners are Leonardo, who wrote for himself in mirrored script, and Lewis Carroll, who wrote mirrored letters to entertain others. Another deliberate mirror-writer is the German artist Kasimir Bordihn, whom we have studied, and discussed in our previous article.

Have you been surprised by your article's popularity?

We thought that mirror-writing was interesting, but we weren't expecting this level of response from the public. We would like to imagine that it reflects the quality of our article, but we suspect an interaction of two factors: mirror-writing is quite common, and there is very little accessible scientific information on the subject. Anyone typing mirror-writing into a search engine is likely to end up at The Psychologist website, reading our article.

One consequence is that, for several years, we have had a steady stream of emails on the topic. A few are from concerned parents, seeking reassurance about their child's development, but most are from people who want to tell us about their own mirror-writing. They might have mirror-written deliberately for years, or they might have recently discovered the ability, and be curious to know what it signifies about their brain. We have been pleased to see that some of our work, some of the time, has such a long reach.

Why do you think that mirror-writing is so interesting?

Mirror-writing has intrigued and unsettled humans for centuries, and has been accorded social significance in different ways. It has been reviled as the writing of the Devil's hand, as evidence of witchcraft and, in more modern times, as a marker of mental deficiency in children (this is now fully debunked, by the way). It has amassed cultural weight as a recurrent motif in literature and cinema, favoured for its evocation of a sinister yet familiar reality through the looking glass. Perhaps because of its occult associations, mirror-writing is often regarded by scientists as an anecdotal curiosity, rather than as something worthy of serious investigation; and funding bodies may tend to take the same attitude. But we think that it is a phenomenon of considerable interest, which intersects with fundamental questions about the neural representations for reading and writing, and for object recognition and purposeful action more generally.

How can we account for mirror-writing?

That depends on whether we are trying to explain mirror-writing in children learning to read and write, or in already-literate adults. Mirror-writing arises during childhood because of an innate tendency for the brain to generalise across mirror-image forms and actions. For most objects we encounter, and actions that we learn, the mirror-image versions are equivalent. We should recognise a dog as the same animal, regardless of whether it is facing left or right. The cultural symbols of written language are notable exceptions, because their identity depends on the way that they face: d and b are not the same (as everydoby must learn).

If object recognition systems in the human brain are predisposed to mirror-generalise, then it will take longer, and require explicit instruction, for children to master this exceptional class of orientation-dependent objects. This creates a developmental window within which the abstract shapes of letters are known better than their specific orientations, and mirror-writing will then emerge.

The situation is very different in adults, because written language is overlearned, and writing actions are engrained patterns carried out habitually by the dominant hand (usually the right). Unlike in children, mirror-writing in adults is strongly associated with the use of the non-dominant hand (usually the left).

The 'unthinking' execution of the engrained right-hand action, when transferred to the left hand, is the mirror image movement. Most people can anticipate this, and make a cognitive effort to adapt the action so the word comes out correctly on the page, but factors like brain damage, or psychological stress, reduce the likelihood of such cognitive supervision. The main reason that left rather than right brain-damage tends to cause mirror-writing may be that it often forces a person to write with the left hand, due to paralysis of the right side.

Even when adults deliberately mirror write, it is usually with the left hand, suggesting that they are harnessing the 'unthinking' script of this hand. The adult phenomenon of mirror-writing may thus reflect the (accidental or deliberate) mirroring of a learned action when transferred between the hands, whereas the childhood form reflects an automatic generalisation across mirror-image forms, prior to internalising the culturally-defined directional requirements of written language.

So, would it be normal for children to mirror-write?

Mirror-writing, though striking to see, is an absolutely normal occurrence when learning to write. It would surprise us if there are any children who never make at least some mirror reversals. Rather than being regarded as a mistake, mirror-writing can be viewed as an impressive act of generalisation from a child, who is able to produce mirrored forms that they have never been taught. Parents with a young child who mirror reverses letters or words should enjoy the variety, and should not worry. Such reversals would only be of concern if they persisted well beyond the age by which most children have securely learned letter direction (7-8 years), in which case they would be part of a broader profile of slow literacy development.

Is there any update on mirror-writing from your own research or other sources?

Mirror-writing is still a 'niche' research topic, but a few recent papers have been published on developmental mirror-writing. Jean Paul-Fischer's group in Lorraine (France) had previously shown that children learning a dextrad (left-to-right) language like French (or English) are much more likely to reverse characters that face to the left (like j, z, or 3) than those that face to the right (like k, s, or 6). They inferred that the child may implicitly learn that most letters they see face to the right, and then over-apply this rule, so that they are more likely to flip a left-facing character to the right than vice-versa. We recently confirmed that this bias really is driven by character orientation, and not by differences in frequency, or how hard it is to remember certain shapes. We taught primary school children to write four novel pseudo-letters, two of which were left-facing and two of which were right-facing. We used identical but mirror reflected character sets for different groups of children, to control for any incidental differences between the shapes. Children were three times more likely to mirror-write a novel character they had learned in a left-facing format than to mirror-write one they had learned in a right-facing format.

Interestingly, it turns out that the bias may not be so much about whether the character faces left or right, but whether it faces in the direction of writing. Fischer and his colleagues used a simple technique to bias children to start writing in a right-to-left (i.e. reversed) direction, and they found that the pattern of reversals was also reversed, so that right-facing letters were now more likely to be flipped than left-facing characters. So, it seems that children may generally learn to face characters in the direction of writing before they know which way each of the individual letters should face.

And there is one point in our previous article that we would now revise. We suggested that mirror-writing in children was driven mainly by uncertainty about the direction of writing actions, and not by perceptual uncertainty about how the letters should look on the page. We have now tested this idea more directly, and found that there is in fact a close relationship between a child's likelihood of mirror-writing and the errors they make when perceptually judging whether normal and reversed characters look correct or not. This relationship was significant even when controlling for age; and the letters that were most often mirror-written were also more prone to recognition errors. These new data indicate that perceptual uncertainty does accompany mirror-writing in children, and that visual and motor representations of letters develop in parallel.

What questions on mirror-writing are still unanswered?

One major shortcoming is that most of what we know about mirror-writing relates to dextrad (left-to-right) languages based on the Latin alphabet, which is only one class of directional writing system, so cross-cultural studies seem essential. How do these phenomena compare in other language systems, especially sinistrad (right-to-left) written languages such as Arabic or Hebrew? Bilingual children, being schooled both in dextrad and sinistrad languages, might be particularly interesting to study. We have unpublished data suggesting that children learning to read and write both English and Arabic make more orientation errors for left-facing characters in English and for right-facing characters in Arabic, consistent with a general bias to prefer letters that face in the script direction. It might also be interesting to examine the relation of reading and writing to other culturally-specified directional behaviours (such as turning taps or screws).

In adults, we would be interested to investigate a possible association of mirror-writing ability with atypical language dominance. We have functional magnetic resonance imaging data showing an unusual pattern of bilateral language representation in a skilled mirror writer. This result is intriguing, but it is not yet known whether it is typical of people who have a facility for mirror-writing. The extensive email correspondence that our Psychologist article has elicited has convinced us that there would be plenty of candidates for a larger-scale study. However, in pursuing this question it would be essential to define more precisely what should qualify a person as being a 'natural' mirror-writer; because mirror-writing is also a skill, like any other, that can be developed and made automatic through practice.

Are there other behaviours or phenomena that are perhaps not as rare as we might think, which people may often be Googling for reliable information? Sleep paralysis springs to mind.

Yes, sleep paralysis is a good example. It happens when a person wakes up, but is unable to move. It's a terrifying experience, often accompanied by hallucinations, or the feeling of being at the hands of a malevolent intruder, like a demon or an alien creature. It is not a dangerous state, but those who experience it would often remember it as haunting and fearful, and might attribute it to malign or even paranormal causes, unless they find reliable information to suggest otherwise.

Another phenomenon, which has in the last decade acquired a proper name, is aphantasia: a lack of mental imagery. A well-publicised case of a man who lost his mental imagery following a neurological event, prompted many others to realise that they had had this condition for their whole lives. Like congenital prosopagnosia (face-blindness), which became widely-known a decade or so before, or like synaesthesia, congenital aphantasia is rare, yet there are many such people around (roughly 1 in 50). Increased awareness allows people to identify their experience, and opens up the condition for wider scientific study. At the same time, we should be slightly wary of the legitimacy and implied understanding that quasi-diagnostic labels confer. A fashionable example would be the so-called Impostor Syndrome, in which people doubt that their achievements reflect their true abilities or efforts, and fear being found out as phonies. This may well be a common and interesting phenomenon, but the syndromic label can encourage circular explanations, and may risk conflating, and pathologising, diverse varieties of self-doubt.