Plagiarism and Creativity

To plagiarize is to appropriate ideas, passages etc. from another work or author. I know this because I have just copied this definition from a dictionary (Collins English Dictionary, Third Edition). In the current climate, many university administrators and lecturers would question my claim to know what plagiarism is given that all I have done is copied. How can my understanding be assessed without the production of my own interpretation of plagiarism? Perhaps the remainder of this essay will put their minds at rest.

I teach psychology in a modern university. Over the course of the last year or so the issue of plagiarism has become salient at BSc team meetings. The general view is that plagiarism is increasing due to the huge amount of internet resources available to students as well as the rise of essay writing companies. Senior management worry about the impact of increased copying on the quality of the degree and they assume that there is more copying going on than can actually be detected with current assessment practices. It would seem that quality is in part a measure of the originality of the student body or, rather, that original work determines quality. On the surface this appears reasonable. As a School of Psychology we should want to produce students who can think for themselves and express that creativity in appropriate ways; indeed, this would seem to be a key skill that employers seek. But what interests me is the notion of creativity that is implicitly adopted.

Creativity is generally associated with novelty: the production of something wholly new. A new thing might be a solution to a problem. The problem will have been persistently unsolved for a good deal of time, but perhaps not without many, many attempts. To take a simplified example, heating a room in the winter has been possible since the advent of fire (though, of course, rooms came after fire). If the room overheated, a window could be opened. With the onset of central heating people were less happy to open windows and heat the garden, for the fuel was costing them much more than the energy previously expended collecting driftwood and maintaining a forest. None the less, an overheated room is uncomfortable and possibly detrimental. But what to do? It was not until the invention of the thermostat that heat regulation and energy efficiency could be more or less solved. Bimetallic strips that differentially expanded and flipped switches were the order of the day and costs were reduced.

There are some points to draw from the thermostat example. First, the problem-space was constrained and therefore usefully describable, in this case in terms of heat and energy expense. Second, these constraints focused interest on materials that might be heat sensitive, which could be said to behave or react in a regular way to heat. At root, some sort of information capture was required. If someone came along and suggested that the real solution to the problem lay in the cheese market, as that was the best source of information, they would be rightly judged incompetent. We might laugh at the cheese marketer, enjoying his ludicrous disjunction with the problem at hand but we can only revel in this mirth if we have a good understanding of the problem-space and can make rational excluding judgements. In short, we require a low-level theory of the problem-space to get started.

Problem-spaces vary in size and complexity. The kinds of problem-spaces confronting psychological scientists in their day-to-day research are vast. Much of the work of scientists is developing theories of the problem-space. This is because science is mostly about exploration of how the world is and this is less practical than trying to organise heat regulation in a room. None the less, questions such as ‘how does language develop?’ and ‘what best predicts safe-sex compliance in South African males?’ are fairly straightforward problems with solutions, even if we do not know them yet. When we do know the solutions, something creative and original will have been done.

When you recognize that you are confronting a large problem-space you are already in possession of a theory about that space, for you can define it to some limited extent. Scientists try and produce hypotheses from the initial definition and test them. These hypotheses might very well be highly intuitive and will almost certainly be wrong, but falsifying them adds to the accuracy of excluding judgements, and to knowledge. When the activities of scientists are plotted it will look as if they are taking a random walk through the problem-space: but, such random walks eventually hit upon solutions. Random wandering within a confined space, then, is a route to creativity; and note, that the criterion for creativity is not necessarily located within the individual scientist; creativity is merely a consequence of constrained wandering activity.

Undergraduate students confronted by an assessment, such as an essay, are asked to answer a question. These questions are derived from a curriculum that has been developed within the university, and in the case of my discipline, also by the British Psychological Society. Such assessments do not represent large problem-spaces; in fact, they represent vanishingly small problem-spaces. Schematically all such assessments require a student to identify the correct body of work used in the curriculum (which is a look-up task), to wander around this literature, extract the relevant parts and organise them. For each aspect of the curriculum, wandering paths will be described and criteria of literature organisation related, such that very little-to-no random activity is required. Students are then rank-ordered more or less according to the size of literature space used and various literary judgements about turn of phrase, parsimony and so forth. Clearly, students must understand what they read, such that performance will covary with intellectual measures, but what is being taught is a very limited game.

The aim of the assessment game cannot be to engender originality or creativity but, at best, it must be to point to the parameters of the larger problem-space of scientific psychology. This is useful and will establish knowledge that students can use in the future in order to be creative but, in itself, it is not creative. There is another aspect to the assessment game, and this is to do with literary judgement. Lecturers will pass an essay if it is written in a style that has the property of authenticity. This is often miscategorized as originality; it is not the same thing. Authenticity is a measure of the students’ voice with the assumption that the more of their voice used, the more understanding of the assessment game the student has. So, I have just relayed some of Karl Popper’s ideas in the preceding paragraphs without use of quotation or citation. My expression is authentic and, I trust, denotes some understanding of Popper’s view. For the student to succeed she must be able to convince the lecturer of her authenticity. This is a very human judgement and serves a gate-keeping function for the discipline: only those who convince us that they understand can continue to graduation and maybe beyond.

Assessment of authenticity is what is really challenged by plagiarism. The fear must be that the new technologies, that allow access to more turns of phrase to express the same ideas circulating in the curriculum, will make it harder for lecturers to spot. To resolve this, universities are turning to software that will check essays against vast databases in order to give a measure of authenticity (although, as said, this is not the terminology necessarily used). Lecturers can root out plagiarists and students can measure their own outputs before submission, and tinker with them until their work scores below a threshold of acceptable plagiarism. In the latter case this will help to train authenticity, so some arguments run.

The adoption of anti-plagiarism software is a mistake. If the software is unbeatable, in other words, if it cannot be cheated and is a genuine, fool-proof system it will at some point exhaust all possible turns of phrase for a limited curriculum. Scientific psychology is finite and describable. We might wish to fall back on the infinite possibilities of a generative grammar (which you will note is an oft repeated claim in the psycholinguistics literature) but if we do this then all we will have created is a machine that forces students into increasingly odd linguistic novelties. Effectively, students will begin writing poetry about psychology. At this point it might begin to look creative, but I am not sure that the original task would be addressed any more. Given this, it seems that lecturers, as humans making these judgements, are better suited. Students will learn to express themselves with the appropriate degree of authenticity to get the point across. Every now and then a student will plagiarize and get away with it, but if they can fool the gatekeepers then perhaps the issue is with the latter and what is being kept. An arms race fought between human brains will also have positive effects on the discipline as the gatekeepers will be selected for increasing intellectual virtue.

If the software is fallible (and it will be) a different arms race will be entered into that will drift toward the situation outlined in the preceding paragraph.

Game theory can take us further still. If you want to force students to seriously strive for authenticity then stringent punishments would work well. Currently, students caught plagiarising have mild punishments, but nothing that threatens their overall degree. For some, it will always be worth riding the odds in this situation – occasionally getting caught and losing a few marks versus low maintenance mark production through copying. Replace this with random checks and instant expulsion from the degree and riding the odds will seem like a very risky idea. Most students will opt for attempted authenticity.

The moral of this essay is that universities would do well to reflect on just what exactly they are providing when they educate their students to degree level. This has implications for how we assess our students and how we attempt to regulate their behaviour under the assessment regime. Technological solutions are extremely tempting, not least because they appear labour saving and all-seeing, but their downstream consequences should be modelled and discussed. Human fallibility has its own rewards.