In 1922, F.M. Stenton published one of the most informative and concise introductions to English private charters of the post Conquest period that has ever been written.⁽¹⁾ His contexts were the largely twelfth- and thirteenth-century charters of five Gilbertine houses in Lincolnshire, preserved in early fifteenth-century transcripts on the Exchequer Memoranda Rolls.⁽²⁾ Because only five percent of the nearly 200 charters in his edition were dated precisely to the year in which they were issued, his main concern was to identify characteristics by which earlier documents could be distinguished from later ones. Those characteristics invariably translated into the increasingly formulaic nature of the charter as it developed from the reign of Henry II (1154-89) to that of Henry III (1216-72). The point of departure was the royal writ and the pervasive influence of the royal justice system: ``...the general submission of the greatest men in the land to the ruling of the king's justices upon the phraseology of a charter which is to be [considered] valid is one of the most remarkable results of the work of Henry II and the men who were trained in his court.''<sup>(3)</sup> The ``victory of common form,'' Stenton argued, was ``the victory of the king's judges''.⁽⁴⁾ He was convinced that charter chronology was inextricably tied to the growth and development of formulae in the document.

Typical of the twelfth-century charter, but noticeably less common from the early years of the thirteenth, was the use of specific greetings in contrast to the unaddressed notification;⁽⁵⁾ of third party consent as opposed to first party warranty;⁽⁶⁾ and of vague, in comparison to the progressively more accurate, topographical identification of landed property.⁽⁷⁾ Furthermore, while the wording of twelfth-century charters seldom defined the social status of a tenant, thirteenth-century examples made the distinction between the servile and the free quite clear.⁽⁸⁾ The twelfth-century conveyance treated churches like any other property; that of the thirteenth concentrated on the advowson.⁽⁹⁾ By the second quarter of the thirteenth century, the sealing clause had become a common form of confirmation.⁽¹⁰⁾

Subsequent editors have profited from Stenton's guidance, but few have achieved his scope let alone expanded significantly on his carefully laid foundations. Many disregard the question altogether. Some have reiterated the importance of royal precedents,⁽¹¹⁾ the development from assent to warranty,⁽¹²⁾ the increasing precision in the description of landed property,⁽¹³⁾ and the growing use of the seal for authentication.⁽¹⁴⁾ Others have noted evidence of broader chronological range in charter form. Whereas Stenton associated the ceremonial taking of an oath and pledging of faith with documents of the twelfth century, Philippa Brown has found pledging to be common in the first half of the thirteenth as well.⁽¹⁵⁾ In contrast, it was probably Stenton's statement that grants to religious houses ``in free, pure and perpetual alms'' were a ``perfected formula of the thirteenth century'' that prompted Una Rees to conclude that ``the phrase `in puram' before `perpetuam elemosinam''' in a confirmation of Henry II was ``inconsistent with a date early in the reign''.⁽¹⁶⁾ Further indicators for dating by form have been found in distraint clauses attached to the grant of a rent,⁽¹⁷⁾ in the wording of titles,⁽¹⁸⁾ in a charter's temporal role vis-à-vis the livery of seisin⁽¹⁹⁾ and, from the late thirteenth century, in formulae allowing enfeoffees total freedom over the alienation of property received.⁽²⁰⁾

The importance of the formula as a mechanism for dating medieval charters has nowhere been better expressed than by Wendy Davies in her detailed study of the Llandaff charters: ``It is a commonplace that genuine charters of the same place and period normally use the same or similar formulae to express similar acts and intentions.''<sup>(21)</sup> However, the question invariably arises as to how much other documents have been subjected to what scholars variously describe as ``alteration'', ``editorial contamination'', ``fabrication'', ``rehandling'', ``reworking'', or simple ``forgery''. Davies adds that the problem with corrupt charters in her collection ``is one of explaining both the use of the same formulae in different charters and the very complex pattern which the sum of different usages presents.''⁽²²⁾ There can be no doubt that just as charter formulae change over time and place in response to constant modifications in social circumstances and legal policies, so the content of early charters may have been adjusted by later copyists to conform to current conditions.⁽²³⁾ On the other hand, such emendations are far more likely to occur in papal, royal and episcopal charters which convey privileges than they are in the much more numerous private charters. Several points are to be made in this context. The first is that reworked charters may be considered genuine if they have been formally approved by the grantor, his successors, or a legitimate superior authority, in the form of a confirmation or an inspeximus.⁽²⁴⁾ Secondly, the introduction of new terms or formulae to a given act might lead the unwary historian to misjudge the original date of composition. Yet it has been pointed out that from a statistical viewpoint, such insertions occur too rarely to mask the original syntax.⁽²⁵⁾ Even in cases exhibiting numerous changes, the prototype is seldom obliterated.⁽²⁶⁾ In fact, if the text of a charter is considered as a whole, the date of composition of the different stages may be identified by the currency of the word patterns and formulae appearing in it. Finally, the concern that cartulary copies of charters may have been badly corrupted by scribes during one or a series of subsequent recopyings is probably exaggerated. Generally speaking, professional medieval English scribes working with legal materials in the post-Conquest period were well trained and made accurate reproductions of the texts. It has been shown that even after five recopyings, twelfth-century charters appearing in a fifteenth-century cartulary remained faithful to their original forms.⁽²⁷⁾ Changes in later recopyings tend to consist of emendations to the spellings of place names to conform to contemporary usage, of abbreviations of the terms of a conveyance, and of reductions or even elimination of the witness list.⁽²⁸⁾ While abbreviation will certainly reduce the number of forms available for examination, none of these changes are likely to affect the structure of the forms themselves.

In the absence of precise dates, or datable events, written into the text of a document, historians have had recourse to a variety of internal and external indicators allowing the determination of approximate or circa dates. By far the most common method of internal dating is the association of names in a given undated document with their identifiable counterparts in dated sources.⁽²⁹⁾ Obviously, the greater the number of identifiable names and date ranges available, the greater the potential for determining an accurate date. However, problems do arise. There are cases in which witnesses from older sources have been appended to later lists,⁽³⁰⁾ and there are lists containing names that have exact counterparts in preceding or succeeding generations.⁽³¹⁾ The greatest difficulty of all is that in many cartulary copies, especially the later ones, the witness lists are truncated or omitted altogether.⁽³²⁾ In such situations, the task of assigning anything but broad date ranges has proved to be exceedingly difficult. Occasionally, a terminus post quem for a series of charters can be achieved by determining the date of compilation of a cartulary or its component sections.⁽³³⁾ Finally, in the case of original documents and to some extent even of cartulary copies,⁽³⁴⁾ there is the test of handwriting, but this too has its limitations within the period under consideration: ``The evidence of handwriting will always distinguish a document written under Stephen from one written under John: it will by no means always determine whether a document was written under Henry II or in the earliest years of Henry III.''<sup>(35)</sup> For Stenton, handwriting could provide a valuable supplement to the dating process, but, he argued, ``in a conflict between the evidence of handwriting and that of formulas the evidence of formulas is generally to be preferred.''⁽³⁶⁾

There can be no doubt that in the absence of witness lists, or a sufficient number of identifiable names within the body of a charter, the best alternative for establishing the document's date is a study of its formulae and word patterns. Despite Stenton's insistence on the importance of formulae for this purpose, however, the great majority of the examples he cites do no more than distinguish between the documents of Henry II and Henry III, representing a quarter century at minimum, or more generally between documents of the twelfth and thirteenth centuries. Using all the methods at his disposal, he is in the end obliged to explain ``the wide limits of time which are assigned to charters in [his] book'' by the lack of a date within the text.⁽³⁷⁾ After so much effort to establish valid chronological distinctions among charters that are representative of most medieval English cartularies and collections of original conveyances, it turns out that the methodology is not, in fact, very useful.

When Stenton produced his edition of the Gilbertine charters more than seventy years ago, the dating especially of private charters by the identification of named parties was little more effective than dating by formulae because the pool of names available for comparison in printed sources was relatively small. Since that time, tens of thousands of historical documents have been published, making dating by association of names an increasingly accurate procedure. As a result, attention to formulae for this purpose has largely been considered a last resort when names are either lacking or otherwise unidentifiable. In the many cases where witnesses have been omitted from cartulary copies, editors have been satisfied to ascribe such terms as ``early'', ``middle'' or ``late'' to charters of the twelfth and thirteenth centuries, or to apply date ranges extending over more than half a century.⁽³⁸⁾ One may well question the value to the historical record of a document that is assigned to such a broad space of time.

The reluctance to develop Stenton's examination of charter formulae for dating purposes can be attributed only in part to the increasing availability of identifiable names. It can also be explained by the assumption, promoted by Stenton himself, that formulae, once established, were immutable. He speaks of clerks working out ``a common form of set phrases,'' and concludes by extolling ``the victory of common form ... over the forces of social division which were inherent in feudalism.''⁽³⁹⁾ Writing more than seventy years later, Michael Clanchy refers to the ``set phraseology'' apparent in charters describing the conveyancing of property.<sup>(40)</sup> It is this conviction that has led many English historical societies involved in the publication of records, ostensibly for economic reasons, to encourage their editors to calendar the content of thirteenth-century charters, particularly their formulaic content, if they can be attributed to the period post-1250.<sup>(41)</sup> Some editors calendar from as early as 1200,<sup>(42)</sup> or the end of the reign of King John,<sup>(43)</sup> while still others calendar everything.<sup>(44)</sup> Statements to the effect that ``witness lists and most dating clauses have been given in full'' attempt to reassure the reader;⁽⁴⁵⁾ as though witness lists were infallible.

The practice of calendaring undated English charters deprives the historian of the most secure evidence for dating that such documents can provide, namely vocabulary, phrases, formulae and syntax. Contrary to popular assumption, the forms and phrases of medieval charters were never ``set''. The content of such records certainly became progressively more formulaic during the course of the twelfth and thirteenth centuries, and thereafter, but the formulae exhibited an enormous variety in and of themselves which changed over time.<sup>(46)</sup> Not only did the formulae change, but so also did the syntax in which they were written. Frequently enough, these modifications reflected the social, political and economic context in which a given document was produced, but there is also evidence of stylistic preferences being current at a particular time, not to mention styles which can be attributed to individual scribes. The language, form and content of the medieval charter can thus be seen to be in constant flux, so much so in fact that when looked at as a whole and compared with other similar documents, the make-up of a given charter can provide an accurate indication of its date. It can be shown by means of electronic sorting that each word and string of words in context has its unique chronological pattern or ``fingerprint''. Taken together, a document's fingerprints can determine its date.

Modifications to charter terminology are equally apparent in every part of the document from the address to the sealing clause. Forms of address limited to the twelfth century include `matris ecclesie filiis presentibus et futuris', 1150-95, with one occurrence c. 1235; `notum sit vobis quod ego/nos pro salute mea/nostra', 1155-96; `omnibus hominibus suis et amicis [Francis et Anglicis]', 1165-90; `sancte Dei ecclesie filiis', 1165-1200. Some extend into the following century: `omnibus sancte matris ecclesie', 1150-1265 (fig. 2); `presentibus et futuris salutem', 1145-1230; `sciant tam presentes quam futuri', 1150-1240. The thirteenth century, especially the second quarter, establishes new formulae: `noveritis me', 1210-1375; `omnibus Christi fidelibus', from 1225 (fig. 2); `omnibus hoc scriptum', 1230-80; `universis sancte matris ecclesie filiis ad quos presens scriptum pervenerit', 1225-85. The salutation `salutem in Domino' first occurs c. 1185, but it is rare before c. 1230 and falls off sharply from c. 1260 to a final occurrence in 1377.