Sir William Clarke's Shorthand


Introduction to G. E. Aylmer, Sir William Clarke Manuscripts

Worcester College, Oxford (Harvester Microfilm Publications, 1979)



Much archive material of high potential interest and value has remained unread because it is written in a condensed, concealed or cryptic form. William Clarke's shorthand is a striking example. It was presumably used to record the 1647 debates transcribed in Ms.5.6. If so, that original shorthand text has not survived; but the defects of the longhand version suggest that Clarke's stenographic skills had been rather recently acquired. By 1649 however he was a practised shorthand writer (MS.2/5 (CXIV) f.111) and the samples copiously interspersed throughout the continuous records from 1651 to 1660 (MSS.1/12-1/21 (XXIV-XXXIII)) are fluently cursive yet reasonably distinct in outline. There is no evidence however that any of them had ever been transcribed or read by anyone other than Clarke himself until I did so in 1973.

   Shorthand, like code, cipher and stenography, confronts the historian with unfamiliar and often uncongenial technical problems. Some progress has recently been made in the international coordination and dissemination of cipher-breaking endeavours in the historical field, and in informing historians about basic cryptological techniques, some of which are also relevant to early shorthand systems. Otherwise little attention had been paid to the elucidation of such systems, even though their historical content may well be of more general interest and value than that still concealed by code or cipher.            Their very popularity and proliferation present daunting problems. In the half-century following the first publication of a complete English shorthand alphabet (John Willis, 1602) some 30 others are known to have been devised and developed. All achieved secrecy as well as speed; each could be used personally and idiosyncratically; most had the same basic principles; many adapted or indeed plagiarised the others; some were republished in heavily-revised editions; and at least one was rejected and replaced by its inventor.

   The system used by Clarke answers to each of these descriptions. Not surprisingly, attempts to read it by identifying the system have proved unsuccessful.  A far more promising procedure is the quasi-cryptological method of counting and classification. That approach has two further justifications. First, the early shorthand systems were often quite akin to cipher, not only in providing effective concealment but also in being more orthographic than phonetic and hence to that extent amenable to linguistic analysis such as frequency-count. Further, the cryptological approach works in practice. Thus the Pepys diaries lay unread in Magdalene College Library for a century and a half until their system was effectively decrypted by Lord Grenville (1759-1834).  As a law student he had had some experience of shorthand; as a statesman he knew something of the techniques of cryptanalysis. According to his own account, it was the latter on which he relied; and this alone would have sufficed for success, whereas shorthand experience along, (short of familiarity with the actual system) would not. Once its basic features are established and defined, shorthand can more readily be traced to its published source, which will then in turn reveal more details. Pepys used Thomas Shelton's Tachygraphy; and so did Clarke. Dating and other evidence points to one of the many editions published in the 1640s. By 1650, Shelton, presumably dissatisfied with his own system, had brought out a new one.  Both now seem rather weak as well as derivative; pure chance determined that the elder foundling would be not only widely adopted, but fostered by two of the most prolific record-keepers of the 17th century. At least its ambiguities and other weaknesses as shorthand, such as a tendency to keep symbols disjunct rather than conjunct, render it the more readily vulnerable to analysis, the main points of which (perhaps relevant to other more refractory shorthands) are now briefly summarised.

   A first examination of 15 specimen xeroxed pages showed that the Clarke shorthand contained about as many basic symbols as there are letters of the alphabet. Conclusion: each letter has one basic equivalent, as in simple substitution cipher. Inference: the process of conversion from word to shorthand relies on spelling rather than sounds. (Corollary: those symbols which resemble letters (e.g. X) represent those letters; the system is being used primarily for convenience, not concealment). Further counting and classification demonstrated that certain basic symbols were consistently combined in five main relative positions, thus -




             X 3




where X is any such symbol and positions 1 - 5 illustrate its alignment to any other. Conjecture: this is a way of indicating medial vowels – a-e-i-o-u – in-order 1 - 5. To take an actual example: 1 means B, / means T: so /1 is BAT, 1/ is BET, and so on. Corollaries: symbols thus used stand for consonants, so their relative frequency should correspond to the consonant-count of ordinary language. Similarly e appeared in a context that suggested  'June' (1651); and occurred far less frequently than _. So ex hypothesi c means J; and _ means N; and c means J followed by N in the U-position. Corollary: terminal. 'e' is not expressed; thus 1/ may mean BIT or BITE. Again, a corrected draft passage was headed Pv;ex hypothesi, this is not a conventional abbreviation but a strict application of the rule already described, viz. P means S, v means R, and Pv means S followed by R in the I-position; hence 'Sir' or 'Sire'. Those and analogous inferences were supported by the linguistically\ plausible results of their substitution elsewhere in the shorthand text. Eventually a detailed picture of some 360 equivalents (single letters, prefixes, suffixes, words, etc) was compiled and recorded, in degrees of certainty ranging from tentative conjecture to complete assurance. This in turn enabled the source-book to be identified and consulted; it lists many more such equivalents. Thus Clarke's shorthand is now in principle readable wherever it is legible.

   In this process of analysis, several phrases and sentences were reconstructed; and these offered some pointers to the general content of the shorthand passages. Not surprisingly, they transact the day to day business of an invading and occupying force. One or two of the entries deal with Clarke's personal and family affairs; but their more typical topics are manouevres, skirmishes, inventories of forces and weapons, pay, promotions, postings and punishments, and in general the routine standing orders and correspondence of a military secretariat on active service.  Thus on the page headed Dalkeith Jan 26 1651/2 (MS 1/10 (XXII) f. 15, recto) the first item is a petition of the Right Worshipful Commissioners for the Affairs of Scotland, and the second is a news report about how Highlanders had murdered some of Cromwell's footsoldiers in their beds, and stolen away three score of cattle from a Lowland gentleman. The notes from Edinburgh dated 25th June 1651 (MS 1/7 (XIX) f. 40, recto) deal with appointments to commands, and arrears of pay; Major-General Desborough, Lieutenant-Colonel John Read, and Major George Scutt are among those mentioned. The second item recto begins Pv (=Sir) and ends 26° c (=26th June). On the same day (ibid. , verso) evidence in the form of a copy letter dated 1647 from Sir Thomas Fairfax was produced to show that Lt. Col. Read had been appointed 'governor of Poole and Drownsea Castle in the county of Dorset and commander in chief of all the forces raised or to be raised for the defence of the same.' Here finally, to facilitate comparison, is a complete transcription (verifiable by a longhand version of the same text) of the first four lines under the heading Leith Febr. 25 1651/2 (MS 1/10 (XXII) f. 311, recto). 'This day one Wragge who was formerly a sutler in the army having a wife in England and since his coming in Scotland married another here which being discovered with was this day sentenced by a court-martial to be tied down to the gallows and after that to have twenty stripes from the main guard to the sand-port and so turned out of town which sentence was executed this day accordingly.'

   Thus even in its reporting of mundane topics and incidents the shorthand affords lively glimpses of service with a Puritan army. Transcription is of course an arduous and time-consuming task; the shorthand is far from uniformly legible, and it is erratically as well as idiosyncratically used. But although this new mine may not always be easy to work, it is certainly extensive and unexploited. May it also prove profitable.




T. Shelton, Shortwriting (1626), 2/1630 R/1635 as Tachygraphy, 6/1641, 7/1647, etc.

Id., A Tutor to Tachygraphy (1642)

Id., Zeiglographia (1649)

W. Grenville, Letter of 21 August 1818, to the master of Magdalene, Published in

   Illustrated London News, 24 April 1858, p. 407

J. Dailey, 'On the Cipher of Pepys's Diary', Papers of the Manchester Literary ClubII 

   (1876), pp.130-7

T. Anderson, History of Shorthand (1882)

A. Wright, T. Shelton, Tachygrapher (1896)

I. Pitman, A History of Shorthand (4/1918)

W. Matthews, 'Samuel Pepys, Tachygraphist', Modern Language Review XXIX (1934), pp.


Id., Introduction to Shelton's Tutor (1642) and  Tachygraphy (1647) (R/1970)

A. Leighton, 'Coordinated Historical Cryptanalysis: Codebreaking as a Historical Resource',

    Proceedings of  XIIIth International Congress for the History of Science, Moscow 1971 (1974)

Id., 'Progress Report on Historical Cryptanalysis', Proceedings of the XIVth International

   Congress for the  History  of Science, Tokyo 1974 (1975)

Id., Some Examples of Historical Cryptanalysis', Historia Mathematica 4 (1977), pp. 319-337

E. Sams, J. Moore, 'Cryptanalysis and Historical Research', Times Literary Supplement (4 March