Harold Bloom famously dubbed it the \"anxiety of influence\" (paywall): the effect which the literary canon has on writers. Less today than it did in the past, according to a mathematical study which analysed thousands of works written over the last 500 years. American mathematicians, led by the chair of Dartmouth College mathematics department Professor Daniel Rockmore, set out to investigate \"large-scale\" trends in literary style. Using digitised works in the Project Gutenberg library, they processed 7,733 works from 537 authors written after the year 1550, were looking for the frequency at which 307 \"content-free\" words – such as \"of\", \"at\" and \"by\" – appeared. They called these words the \"syntactic glue\" of language: \"words that carry little meaning on their own but form the bridge between words that convey meaning\", and thus \"provide a useful stylistic fingerprint\" for authorship. \"When we consider content-free word frequencies from a large number of authors and works over a long period of time, we can ask questions related to temporal trends in similarity\", they write in their new paper, \"Quantitative patterns of stylistic influence in the evolution of literature\". After finding that authors of any given period are stylistically similar to their contemporaries, they also discovered that the stylistic influence of the past is decreasing. While authors in the 18th and 19th centuries are still influenced by previous centuries, authors writing in the late 20th century are instead \"strongly influenced\" by writers from their own decade. \"The so-called \'anxiety of influence\', whereby authors are understood in terms of their response to canonical precursors, is becoming an \'anxiety of impotence\', in which the past exerts a diminishing stylistic influence on the present,\" they write. This could, they suggest, be explained by the modernist movement, in which authors \"reject their immediate stylistic predecessors yet remain a part of a dominant movement that included many of their contemporaries\". There are also, they point out, more books available to read these days. \"One hypothesis is that there is so much more to read now and more kinds of \'important\' work that if we believe that style is influenced by what one reads, then it is less likely that people generally devote the preponderance of their reading to the older \'classics\',\" said Rockmore. \"This is not unrelated to the acknowledged gradual diminishment of a canon in literature. If one believes that writing style is significantly related to spoken language, then it might be traced to the rapid evolution of that form of communication.\" The paper only covers works written up to 1952, because of copyright issues, but Rockmore believes the decreasing influence of the canon will only have continued in the authors of today. The Orange prize-winning novelist Lionel Shriver agreed this was probably true in her case – and suggested it was likely to apply to her contemporaries as well. \"About all I can do is confess that while I myself devoured classics in my teens and 20s – even 30s, come to think of it – I now read contemporary fiction almost exclusively,\" she said. \"I feel ambivalent about this evolution, but between reviewing, blurbing occasionally, and keeping up with what\'s out there on general principle I don\'t often get around to touching base with the literary canon. When I have tried to, say, reread a Dostoevsky novel, I\'ve discovered that I don\'t have the patience any longer – for the long philosophical digressions, for example. I bet I\'m not alone in this reduced tolerance for the stylistic traditions of the past.\" Rockmore said he hoped to talk to literary colleagues to investigate if the mathematicians\' formulae could be applied to literary criticism. \"I do believe it is an interesting approach to literary analysis,\" he said. \"We are hopeful that we can engage some colleagues expert in those areas to find what the interesting questions are. That is a necessary first step and we look forward to collaborations to those ends.\"