Academic authors in the twenty-first century have become used to submitting an ‘abstract’ of their paper alongside the full text – but abstracts were originally something written by another person.
The practice of ‘abstracts’ arose from a recognition of the value of providing short summaries of a paper, for the benefit of those people who were not able to access the full original. For instance, in the late 18th century, the Royal Society used the term ‘abstract’ to describe the summary of a paper that was written into the minute-books by the secretary after a paper had been read out loud at a meeting. And in the 19th century, some entrepreneurs created journals whose purpose was to provide readers with summaries of articles published in other journals, particularly those in foreign languages. For instance, by the mid-19th century, German publishers and societies had taken the lead in surveying and summarising the latest international papers in disciplines from geology to zoology for German-speaking readers (see Manzer, 1977). The (London) Physical Society had a similar aim in mind when it launched Abstracts of Physical Papers in 1895 (later, Science Abstracts): to provide access to foreign-language material.
Notably, in both these cases, the summary was written by a third-party, and was intended to stand in place of the full paper (rather than appearing alongside it).
At the Royal Society, the secretary’s ‘abstracts’ were used in the editorial process for the Philosophical Transactions: until the 1830s, the editorial committee usually worked from the abstracts, only rarely scrutinising the full manuscript. (This changed with the introduction of refereeing, see Moxham & Fyfe, 2018)
Being shorter than the original, abstracts took up less space in the minute-books, and were quicker to read – they were also quicker and cheaper to get into print. This would give them an important role in enabling rapid, if brief, communication of scientific results.
In the nineteenth century, the Transactions was being published just twice a year, and by the 1820s, some people already thought that was too slow. After all, the Society was hearing papers at meetings held every week, and other publishers were issuing newer journals on a monthly schedule.
So, in 1830, the Royal Society decided to make more use of the abstracts that its secretaries were still writing into the minute-books. It agreed to let its printer, Richard Taylor, use them as the basis for reports on Society meetings in his own journal, the (monthly) Philosophical Magazine; and it also agreed that he should print the secretaries’ abstracts in a separate publication known as the Proceedings of the Royal Society, which first appeared in late February 1831. It also recognised the value of these abstracts as a convenient reference resource by publishing a retrospective collection of the abstracts from 1800-30. (For more on Proceedings, see Csiszar, 2020)
The Proceedings was initially intended for internal distribution to the fellowship, but it was soon put on public sale. Through the nineteenth century, it continued to appear roughly monthly, carrying the secretaries’ summaries of the papers presented at the Society’s meetings. A selection of those papers, as decided by the refereeing and editorial process, would later appear in full in the Transactions. The full papers could be 20, 30 or 40 pages in length; the abstracts were shorter, but not as short as we might expect today – some ran to 4 or 5 pages!
The first explicit evidence we have of author-generated abstracts comes from 1892, when the Society sought to relieve the burden on its secretaries by asking authors to supply an ‘abstract’ for publication in Proceedings at the same time as submitting their paper to Transactions. There is, however, circumstantial evidence suggesting that this was already being expected in the mid-1870s.
Author-generated abstracts continued to be used – as the secretarial summaries had been – for various purposes within the Royal Society (not just for publication). In the 1890s and early 20th century, they were circulated to attendees ahead of Society meetings at which the paper would be orally presented. By the 1930s, they were being used to ‘expedite’ the refereeing process (perhaps by sending them with the invitation to potential referees?).
When the Proceedings had been re-launched as an independent journal in 1905, it no longer published the abstracts of papers submitted to the Royal Society. These now appeared in a separate insert, though that practice ceased during the Second World War.
After the war, the Society began to publish the abstract at the start of the full article (rather than publishing in separately, in advance). It was also around this time that we find more attention – in the Society and more broadly in the world of science – to the abstract, and the ways it could be used to improve scientific communication.
Standardisation of abstracts
The first evidence we’ve found of efforts to standardise abstracts comes from 1914, when it was requested that every paper communicated to the Society:
‘must be accompanied by a summary not exceeding 300 words in length, showing the general scope of the communication, and indicating points which, in the opinion of the Author, are of special importance.’ [revised standing orders, in RS Council Minutes, 21 May 1914, para 37]
That provision was still being quoted to prospective authors in the early 1930s [e.g. Memo, November 1933, in RS archives HD/6/10/2/4]
In 1949, the Royal Society sponsored the preparation and circulation of a ‘Guide for the Preparation of Synopses’, to encourage other learned society publishers to adopt the practice. It recommended that a good synopsis should be no longer than 200 words, written in ‘normal rather than abbreviated English’, and ‘should be intelligible in itself without reference to the paper’. (It is not clear quite what the choice of ‘synopsis’, in preference to ‘abstract’ – or the older ‘summary’ – was intended to convey; but ‘synopsis’ appears to be a subset of the genre of ‘abstract’)
Amidst the growth of scientific research after the Second World War, abstracts (and indexes) were seen as important tools to help researchers navigate a vast flood of literature that was too much to read. When, in 1948, the Royal Society hosted an international Scientific Information Conference, indexes and abstracts were discussed as much, if not more, than research journals themselves. A similar conference was held in Washington in 1956. Developing a more effective system for abstracts was agreed to be key – but it did depend on what the purpose of abstracts was supposed to be, and for whom…
The purpose of abstracts?
In 1949, an article in Physics Today (Gray, 1949) reported preliminary work being undertaken by the American Institute of Physics and the American Physical Society to investigate how researchers actually used abstracts.
The article noted that abstracts served various purposes – for spreading ‘current awareness’, for later reference, or to help with the compilation of indexes – and it noted a range of styles, from the short ‘indicative’ abstract (largely intended to help readers decide whether to read the full article), to the longer, more ‘informative’ abstract that ‘summarizes the article’s major arguments and gives the principal data and conclusions’ (and that, for many readers, could stand in place of reading the full article).
Some commentators had high hopes of a well-written abstract. In 1938, Royal Society president William H. Bragg had argued that such abstracts had a particular value in a world in which scientific research was becoming increasingly specialised and technical:
‘The ideal summary is more than a mere digest or shortened form of the paper. It differs from the paper itself in that it is addressed to a wider circle of readers which may include the experts but contains also many other who should, in fact, receive the principal attention. For this reason it may be more difficult to write than the paper itself…’ (Bragg, 1938, p.17).
Who should write abstracts?
The Royal Society asked for author-generated abstracts, but not all journals yet did this; and there was still a strong tradition of dedicated abstract journals and agencies producing third-party summaries of recent work, usually focusing on particular fields. According to Physics Today:
‘There is great difference of opinion… concerning the quality of abstracting that can be expected from authors. It is argued on the one hand that the author is by far the best judge of his article and of the significance of its contents and, therefore, is obviously the best m a n to write the abstract. On the other hand, it is claimed that the author’s very closeness to the work described makes it impossible for him to view his article in proper perspective and write an adequate abstract.’ (p.22)
Third-party summaries might be better-written and more widely-comprehensible, but they did take time and money to produce, usually appearing retrospectively. This was why the Royal Society’s 1948 conference supported the wider use of author-generated abstracts, despite their disadvantages:
‘The present general unsuitability of authors’ summaries for use as abstracts is recognised; nevertheless, if these could be used it would increase the speed of publication and reduce the cost of journals publishing abstracts.’ (Circular letter from Royal Society to UK learned societies, 12 April 1949)
There was also much discussion around this time about whether microfilm, punched cards or machine-sorting might transform the ‘scientific information system’. In 1949, Physics Today thought that such ‘push-button library research’ was ‘hardly around the corner’. Converting scientific papers, or their abstracts, onto punched cards was certainly a non-trivial challenge at a time of ever-growing research output. The technology and finances required for such work proved to be beyond the capacity of the learned societies and academic institutions that had supported similar bibliographic efforts in the nineteenth century. It thus fell to specialist bibliographic organisations, such as Eugene Garfield’s Institute of Scientific Information (founded 1956), to provide these services in the 1960s and 1970s.
The ease with which author-generated abstracts can be circulated digitally and globally in the early twenty-first century would have amazed the participants at the Royal Society Scientific Information Conference in 1948, let alone the Royal Society fellows who agreed to publish abstracts in the Proceedings in 1830.
But even so, we might note that not all journal publishers do make their authors’ abstracts freely available, something that the Initiative for Open Abstracts is currently seeking to address. The general comprehensibility of author-generated abstracts also continues to be a concern, leading some journals and organisations to request authors to provide ‘plain language summaries’.
William Bragg’s ambition that well-written abstracts might help researchers from other fields grasp the key features of the latest research has not come to pass. Rather, that function seems to have been transferred to review articles, with their aim to summarise and evaluate not just one article but all the latest contributions to a field. Bragg would no doubt be pleased that publishing review articles has been one of the functions of Philosophical Transactions since the late twentieth century.