Why Graphology is a Pseudoscience

What is Graphology

The International Graphonomics Society (IGS) defines graphology as “…the study of character or personality through handwriting” (Simner & Goffin, 2003). The exact time that graphology came about is unknown, but it may have started out in China. Between 100 and 1000 AD, Chinese emperors used a form of graphology to choose priests, politicians, and emissaries (Graphocentric, 2011). Camillo Baldi started the practice itself in the 17th century, and the term was coined in 1897 by the father of modern graphology, Jean-Hippolyte Michon (Oliveira, L., Justino, E., Freitas, C., & Sabourin, R., 2005). The theory was put forward in his graphological journal, which he created after an unpleasant partnership with a chirologist who was also working on graphology (Landau, 2007). One of his pupils, Jean Crepieux-Jamin, continued his work, and improved the organization of the signs used by graphologists (Graphocentric, 2011). The classifications he came up with are still in use today. There are many different types of graphology which are currently in practice, but they commonly make predictions based on each individual letter as well as larger groupings (Simner & Goffin, 2003). Many factors are included in the predictions such as the shape, size, and orientation of the letters. These parameters are examined to determine the personality traits of the writer.

Theory

Graphologists consider several different features of an individual’s handwriting to determine what personality traits he/she has. These features are called the “signs” of the person’s handwriting (Simner & Goffin, 2003). The order, proportion, dimension, form, pressure, constancy, characteristic gestures, and use of space are all considered (Oliveira et al., 2005). The order defines the organization of the writing (Figure 1a). The relationship between the heights of the letters defines the proportionality of
Signatures.jpg
Figure 1: Signatures detailing different graphological characteristics. (Oliveira et al., 2005)
the writing (Figure 1b). For example, if all of the letters are approximately the same height, then the writing is proportional. The dimension is defined as the ratio of the height to the width of the letters, and has two major end members. These are high-dimension (large ratio) and low-dimension (small ratio) (Figure 1c). The form of the writing is determined by the shape of the letters (eg: they can be large and round, or narrow spikes) (Figure 1d). The pressure is simply the force the writer puts on the writing utensil when he/she is composing the words. As with pressure, characteristic gestures are exactly what they sound like. These encompass everything from the way the writer starts and finishes words and phrases to how he/she crosses T’s. Lastly, the occupation of space defines how the writer makes use of the area given for writing. Other features that are examined are the slant of the letters and the orientation of the words with respect to the horizontal (this is used more commonly in signature analysis).
According to graphological theory, these simple signs, on their own, only present a limited view of the writer’s personality. Complex signs and the law of resultants are much more important (Landau, 2007). It is stated that there are more characteristics than signs, and that each characteristic is described by a set of simple signs. This set is called a complex sign, and this is what is used to complete analyses. The law of resultants states that signs become superimposed on each other as one is writing, creating a compound sign with a different meaning than the simple signs on their own.
Using the above parameters, graphologists claim to be able to determine numerous characteristics associated with an individual (Fluckiger, F., Tripp, C., & Weinberg, G., 1961). These traits include the writers’ interests, intelligence, and many others. For example, the dimension of the writing can show if the individual is generally enthusiastic (Oliveira et al., 2005). Attempts have even been made to produce computer programs to expedite the work of graphologists (Figure 2) (Sheikholeslami, G., Srihari, S., & Govindaraju, V., 1943). These programs have had the same degree of success as a regular graphologist, but produce results more rapidly.
Sample1.jpg

Feature
Trait(s)
Left and right justified
Consistency of work, quick and clear comprehension
Constant distance between lines
Regularity and clever mind
Slants to the right
Extroversion
Large lower zone
Intensely pursues goals

Figure 2: A sample of handwriting and its associated analysis, adapted from Sheikholeslami et al., 1943. This analysis was done using a graphological software package. The software generally produced the same analysis as a graphologist, and uses most of the rules that graphologists use. (Sheikholeslami et al., 1943)

Current Use

The current popularity of graphology is highly variable between countries. In some countries, such as France, it is used regularly, whereas Swedish companies rarely use it (Table 1) (Bradley, 2005). It also varies based on the type of job. For example, consultants use graphology more regularly than other companies. It is sometimes used to confirm impressions about potential employees after an interview. In 1983, it was estimated that 3000 American firms used graphology and that the number was actually increasing (Ben-Shakar, G., Bar-Hillel, M., Bilu, Y., Ben-Abba, E., & Flug, A., 1986). It was also found that 85% of European firms used graphology, and that it is the most extensively used personality test in Israel. Despite the evidence against its accuracy, another article published in 1992 showed that these values were generally the same as those just cited (Simner & Goffin, 2003). It was also found that advanced countries such as Germany and Switzerland are showing the same increase in use as is seen in the US. What is more surprising is that this practice is not limited to minor or obscure companies. For example, Barclay’s, Bank of America, and Federal Express have all allegedly used graphology in the hiring process.
Country
Estimated Number of Graphologists
Active Population (Millions)
France
5000
25.871
Switzerland
500
3.860
Israel
200
2.100
Italy
2000
22.680
Netherlands
500
7.358
Hungary
200
4.095
Spain
700
15.625
Belgium
180
4.237
Denmark
50
2.812
Austria
60
3.881
Germany
600
40.083
Norway
30
2.186
UK
200
28.271
Sweden
30
4.319
South Africa
60
11.624
Japan
20
66.660
Table 1: The estimated number of graphologists in several developed countries (in 2000), adapted from Bradley, 2005. It is arranged in descending order based on the percentage of graphologists in the active population.
Apart from the personality assessment commonly given by graphologists, the signs used to make predictions do have a more technical use. For example, it can be used in signature verification, as individuals have diagnostic pen-strokes when signing documents. Thus, graphological characteristics such as dimension and form can be used to detect forgeries (Oliveira et al., 2005).

But is Graphology a Science

Studies

There are two major types of study in graphology. These are the holistic and atomistic approaches (Fluckiger et al., 1961). Holistic studies are designed to determine if handwriting samples actually do provide information about traits of the writer. Basically, this type of test determines if graphology itself is reliable. An atomistic study, on the other hand, examines particular handwriting features and determines how they relate to certain personality traits. For example, the pressure an individual writes with can be measured, and this can be related to his/her personality type.
Conducting a study in graphology is fairly challenging. In a holistic study, an entire sample of writing is used, and the graphologist must either match the sample to a description of the writer or he/she must rank the sample relative to the other samples based on a particular personality trait. For example, the graphologist may be tasked with ranking individuals based on their masculinity. The matching test suffers from a serious problem. If the graphologist makes a correct match, then the odds of him/her making another correct match increases. The opposite happens when he/she makes an incorrect match. Obviously this will affect the final results, rendering these tests inconclusive. Many graphologists appeal to this sort of evidence, claiming that they can correctly match people to their profession. They also often use client testimonials regarding the success of their predictions. This is another form of the matching test, and should be regarded as inconclusive. The ranking test is also inconclusive. The major problem with this type of study is that there is not an accepted definition of various personality traits. For example, there is no set way to determine if someone is more masculine than someone else, particularly if the gap is fairly small.
Testing individual signs, as is done in atomistic studies, is much more conclusive, but is difficult to carry out. These tests involve making an actual measurement of a certain sign, which could include the pressure applied while writing or the speed at which the individual writes. Graphologists may argue against this type of test as it removes the art of graphology from the work and turns it into a simple mathematical process. As can be seen, the actual studies done, particularly by graphologists have serious problems.

Theoretical Issues

As mentioned earlier, graphologists rely on simple and complex signs as well as resultants to make their predictions. The major issue with this practice is that these signs all overlap, making them difficult to pick out. As a result, one graphologist may put more weight on a certain sign than another would. This leads to inconsistencies between predictions, which make it difficult to make a repeatable test. This gives graphologists an easy way to ignore tests that do not support their views.
Another issue is that everyone is taught to write more or less the same way as a child, so these signs do not show themselves until later in life, even though an individual still has their personality as a child. A person’s occupation will also affect the signs. For example, a court stenographer will likely have clear, regular writing, with efficient use of space, as he/she would need to ensure the work is legible. On the other hand, a doctor who repeatedly writes short prescriptions everyday will likely have large, messy writing. This will obviously impact the results given by the graphologist, and while certain people do tend, for example, to be doctors, the prediction will no longer be based solely on personality. Instead, the writer’s occupation will impact the profession that the graphologist would assign to him/her.

The Facts

Graphologists claim that they do not use the autobiographical material in a sample of writing (Simner & Goffin, 2003). Realistically, whether or not they use that material should not have an impact on their final conclusions. This is because the conclusions are based on the signs in the writing, which are present regardless of the content of the writing sample they obtain. Nonetheless, the mean correlation between the graphologists’ analyses and the actual supervisor ratings of autobiographical material was 0.16, and that of the non-autobiographical material was 0.09 (both of which are quite low) (Table 2). Clearly the analyses completed are significantly influenced by the content of the material. This indicates that the signs employed by graphologists are not really diagnostic.
Stats.jpg
Table 2: Correlations between handwriting analyses performed by graphologists and the related performance rating given by the individual’s supervisor. Note that autobiographical material includes information about the writer, while non-autobiographical does not. (Simner & Goffin, 2003)

Why Graphology is a Pseudoscience

Many arguments can be leveled against graphology’s claim to scientific status. Unfortunately, several arguments, while making factual claims, can be easily discounted. For example, one cannot simply argue that there is no natural basis for graphology, as graphologists will claim that an individual’s psyche influences the signs that are present in his/her writing. Even if one could make this argument, it would not be sufficient proof that graphology is a pseudoscience. While graphologists cannot explain why a connection between writing style and personality type would exist, this is not enough to discount it as a science. For example, no one could explain the cause of continental drift, but it eventually turned out to be true (Thagard, 1978). This is not to say that graphology is comparable to continental drift, but that the argument against it is not strong. Some individuals believe in graphology because it supports the idea that their actions are beyond their control, not because it has a scientific basis. Appealing to this fact is also fruitless. Instead, to show that a discipline is a pseudoscience, one must determine if it passes several criteria. Is it falsifiable or can it be confirmed through experiment? What method of explanation does it employ? Have there been any advances in the theory, or is it being left alone while other theories become more successful? Do the proponents acknowledge its flaws and evidence against this theory, or do they just ignore unwanted evidence?

Verifiability and Falsifiability

A theory is verifiable if one can deduce observation statements from it (Thagard, 1978). Clearly, one can extract observation statements from a graphologist, just by getting one to make a prediction, so it meets this criterion.
Graphology is verifiable, but is it falsifiable? In short, graphology is, in fact, falsifiable. One can come up with tests which could potentially prove it wrong (of course there are the issues in testing graphology mentioned above, but in this matter they are not important). All a researcher has to do is get several graphologists to analyze samples of handwriting and then compare their conclusions to reality. This has been carried out by several individuals, and the results suggest that there is no reason to use graphology (Simner & Goffin, 2003) (Table 2). At best, the results are slightly better than chance. Unfortunately, in light of these results, graphologists have an easy out. They can simply say that the graphologists used in the study were not trained properly or that they did not represent graphologists as a whole. This is aided by the fact that the theory behind graphology is not concrete; it instead relies on an individual’s judgment. Thus, graphological predictions are subject to the errors commonly associated with scales based on human judgment instead of mathematical laws, making graphologists no more exact than figure skating judges. While it has been found that graphologists generally agree on particular interpretations, there is some variation between analyses due to the experience level of the graphologist, how he/she combines the signs, and the level of importance assigned to each sign. Thus, while graphology is falsifiable in theory, it is not in practice.

From Thagard on Astrology

As it appears that graphology is both verifiable and falsifiable in theory, it is apparent that some other method of demarcation must be used to define it as a pseudoscience. In his paper, Why Astrology is a Pseudoscience, Paul Thagard puts forward a principle of demarcation which is very applicable to graphology due to the similarities between the two pseudosciences. The principle is formulated as:
“A theory or discipline which purports to be scientific is pseudoscientific if and only if:

  1. it has been less progressive than alternative theories over a long period of time, and faces many unsolved problems; but
  2. the community of practitioners makes little attempt to develop the theory towards solutions of the problems, shows no concern for attempts to evaluate the theory in relation to others, and is selective in considering confirmations and disconfirmations.” (Thagard, 1978)

Graphologists today still use the signs developed at the inception of this practice (Graphocentric, 2011). Essentially, it is the same practice as that worked out by Jean-Hippolyte Michon. Thus, it has not been very progressive. In addition, it proved to be very difficult to find any studies conducted by graphologists that did not rely on anecdotal evidence or poorly designed and controlled experiments. Furthermore, other types of personality assessments, such as biodata and personality tests, are significantly more accurate than graphological predictions. The correlation between prediction and reality for these tests ranges from 0.31 to 0.41, which is much greater than the 0.09 and 0.16 cited earlier for graphological predictions (Simner & Goffin, 2003). These assessments are also more progressive, as an inaccurate test will be discarded, while a graphologist will likely make the same predictions based on certain signs, even if he/she made an incorrect prediction based on that sign. It appears that graphologists generally ignore studies which refute its validity, such as the one mentioned earlier, and use anecdotal statements to support the theory. To sum up, graphology has been less progressive than personality and biodata tests, is very inaccurate, and graphologists do not improve the theory or relate it to others. They also are selective when considering evidence for or against their practice. Thus, graphology does not satisfy either of these criterions.

The Pros and Cons of Graphology

With all of this information about graphology, is it actually harmful to anyone? Anyone who lost a job opportunity based on a graphologist’s suggestion will almost certainly say yes. Furthermore, this time wasting test could be done more cheaply and accurately using simple personality or biodata tests. Thus, it hurts both individuals looking for a job, and those hiring them.
On the other hand, there is a benefit to using graphological signs. An individual’s writing does not change dramatically over the course of his or her life (Fluckiger et al., 1961). Thus, the graphological signs associated with an individual are diagnostic of his/her handwriting. These signs can then be used in forensics to aid law enforcement by doing things such as detecting forgeries. While the signs used by graphologists can therefore be beneficial, graphology itself is still not useful.

Conclusion

It is fairly widely accepted among philosophers and scientists that graphology is a pseudoscience. It is not easily grouped in that classification, as it meets some demarcation criterion. It puts forward testable predictions which are correct from time to time and it is possible, in theory, to falsify it. Of course, graphologists will hold up the few correct predictions as proof that graphology is in fact a science, but the weight of the evidence against it is more than a few correct predictions can challenge. The theory itself appears to be reasonable at first glance. It almost seems logical that one’s personality would come through in his or her writing. It is likely that it is these two features of graphology that are the reasons why so many people believe in its validity. Unfortunately, it seems to do more harm than help, as countless people have lost job opportunities and wasted money because of the claims made by a graphologist. Ultimately, proving that graphology is pseudoscientific came down to two criterion put forward by Paul Thagard who experienced a similar problem proving the illegitimacy of astrology. Graphology failed to meet either of his criterions. In addition to failing his test, the theory behind graphology is not solid. It is difficult to make repeatable tests due to the nature of the theory. Thus, because it failed Thagard’s tests and is based on faulty groundwork, graphology is a pseudoscience.

References

Ben-Shakar, G., Bar-Hillel, M., Bilu, Y., Ben-Abba, E., & Flug, A. (1986). Retrieved from http://test.scripts.psu.edu/users/k/r/krm10/PSY597SP07/BenShakhar%20graphology.pdf

Fluckiger, F., Tripp, C., & Weinberg, G. (1961). A Review of Experimental Research in Graphology, 1933-1960. Retrieved from http://www.amsciepub.com/doi/pdf/10.2466/pms.1961.12.1.67

Graphocentric (2011, Jan. 3). A Brief History of Grapholog.Retrieved from
http://www.graphocentric.co.uk/graphology-resources/a-brief-history-of-graphology/

Landau, S. (2007). Michon and the Birth of Scientific Graphology. Retrieved from http://www.britishgraphology.org/wp-content/uploads/2012/02/MichonAndTheBirthOfScientificGraphology.pdf

Oliveira, L., Justino, E., Freitas, C., & Sabourin, R. (2005). The Graphology Applied to Signature Verification. Retrieved from http://www.inf.ufpr.br/lesoliveira/download/IGS2005.pdf

Sheikholeslami, G., Srihari, S., & Govindaraju, V. (2003). Computer aided Graphology.

Simner, M., & Goffin, R. (2003). A Position Statement by the International Graphonomics Society of the Use of Graphology in Personnel Selection Testing. International Graphonomics Society. Retrieved from http://www.tandfonline.com/doi/pdf/10.1207/S15327574IJT0304_4