Periodic Reporting for period 3 - MindBendingGrammars (Mind-Bending Grammars: The dynamics of correlated multiple grammatical changes in Early Modern English writers)
Reporting period: 2018-09-01 to 2020-02-29
Our brains have their own rules of economy. We are all creatures of habit. With every repetition they become even more tenacious, so the older we grow, the harder it gets to get rid of them. Yet, paradoxically, routine also helps to innovate, as long as the innovation is somehow connected to it. It's the same way that groundbreaking art by people like Picasso is based on a thorough basis of classical training. In language, a familiar instance of this kind of innovation is that of blends like Brexit. Brexit is intelligible precisely because it is built on existing and familiar words. We learn new coinings like these with little effort and well into old age. However, to what extent similar things happen in grammar is far less known. And yet grammar changes too. Language users (subconsciously) prefer grammatical patterns that are somehow similar or familiar, which because of their familiarity can be stored more easily in memory. So far such processes of change have only been researched at the level of the language as a monolithic object. For instance, we know that in the case of passives that end with a preposition, complex patterns like he was made fun of come historically later than simpler ones such as he was sent for. What we don’t know is to what extent such innovations are picked up across the lifespan of individual speakers. Research on individual speakers has so far focused on intergenerational differences, but hardly on longitudinal change within individuals.
The economical quest for similarity goes beyond building single routines, as similarities are also looked for across different structures. Once linked through similarity, the structures may become even more similar and converge. A recent example in English is the emergence of forms like I’m gonna, I gotta and I hafta, whose similar uses are probably an important reason why they now also sound so similar.
Besides brain economy, language habits are also shaped by social interactions. We tend to pick up language behavior from close contacts or imitate people we admire. This is a major way in which innovative language use spreads through a community. However, such novel language use is at odds with our current habits. Together with age, social distance from the innovator, and personality, motivations of economy will then either be so strong that the speaker does not change their habits, or social motivations win out, and the innovation will be adopted.
In addition to getting at grips with the interaction between generational and lifespan change, the project subscribes to a recent interest in viewing aging as a more constructive process, in line with for instance Michael Ramscar and colleagues’ paper The myth of cognitive decline in the leading cognitive science journal Topics in Cognitive Science. In this article it is argued that the slower reaction times of elderly people are not merely the effect of cognitive decline, but also reflect a huge increase in experience: the older you are, the more (subconscious) choices you have to make when retrieving information from memory. Similarly, the insights that will come forth from Mind-Bending Grammars may contribute to a more balanced view of what aging does to the brain, and how the power to adapt is not so much lost, as that it changes in nature. In a society where elderly people make up an increasing part of our population, a more balanced view is needed most dearly, as this might ultimately also improve the way we treat our own and other people’s process of aging.
Next to the emergence of new structures, robust shifts have also been found in structures with two well-established different functions, where their relative weight gradually shifts across the lifespan, in line with a trend in the larger community. Where both uses are already known to language users, and the change only affects the respective frequency of their use, no clear age restrictions have been found.
On a more general level, the research has shown that there is no such thing as the average language user. Instead, interindividual differences may be bigger than commonly assumed. Apart from random differences attributable to style and personality, we also were able to corroborate past research that observes a direct relation between innovative behavior and social networks or demography: people who live in a large city tend to have a social network with more weak ties. Such people are exposed to a greater variety of linguistic behaviors and as a result are less averse to adopting novelties. In our data we see a direct correlation between degree of innovative behavior and time spent in London.
The most important methodological enhancement so far has been the development of a method to measure to which degree a certain structure exhibits behavior that is typical of a language’s grammar, and goes beyond the sum of the meaningful words contained in that structure. The future meaning of this is going to be great resides holistically in this utterance and cannot be attributed to a single word, making the structure more grammatical than I am going to buy meat, which is to a higher extent composed of meaningful parts (since going can still refer to motion here).
In order to be able to investigate the issue of how grammatical innovation and its consolidation emerges through the interaction of generational change and lifespan change, big data are needed at the individual level, which allow detailed tracing of linguistic behavior throughout the years. Mind-Bending Grammars has compiled a corpus of such big data, called Early Modern Multiloquent Authors (EMMA) (www.uantwerpen.be/en/projects/mind-bending-grammars/emma-corpus/). The corpus consists of the published materials of 50 prolific authors. The average sample for each author consists of about 2 million words spread over an average career of 36 years. In addition to their publications, the corpus is also linked to a database with biographical information on each of the authors. With its unprecedented per individual sample size, EMMA can be used for in-depth analysis of linguistic behavior in a representative sample of the London upper-class.
To facilitate linguistic analysis an innovative query and annotation tool was developed alongside with the corpus, called Cosycat (Collaborative Synchronized Corpus Annotation Tool), with improved synchronization between researchers’ activities as compared to alternative annotation tools.
Our results so far, including the observation of innovation past adolescence, the effect of social inertia, and the increased realization of innovative potential as the impact of conservative speakers fades out, provide novel empirical evidence that help modelling particularly the weight of the many factors that are involved in the complex process of language change. So far models that try to combine these factors (cognitive, social, demographic) have mostly worked on the basis of computer simulations, but lack robust empirical evidence. In the remainder of the project more results will add to the refinement of this model, and it will also become possible to make the observations more robust by looking into what is shared or not between case studies.