Two Examples of the Impact of Computer Technology on Historical Editing: The Correspondence of William of Orange 1533-1584 snd the Resolutions of the States General 1626-1651

[email protected]

Introduction

The edition and publication of historical documents became a profession in its own right during the nineteenth century, just as the whole of historical science was going through a process of professionalization. History became a science. The quest for historical facts with the help of text critical rules had to contribute to 'objective' knowledge. Critical use of sources became the basis of this new approach. Consequently, there was a demand for reliable texts and, with an eye to the publication of sources, institutes were established and magnificent, bulky series of source editions came to light. The British historian, David Knowles, spoke of 'great historical enterprises' [1]. In Germany, the Monumenta historiae germanica, a series of documentary editions on medieval German history, set the example in 1819. Other series followed, such as the Rolls Series in England, the Collection des documents inédites sur l'histoire de France in France and the Colección de documentos inéditos para la historia de España in Spain. In The Netherlands, the Dutch government established a commission in 1902, now an institute, to publish the Rijks Geschiedkundige Publicatiën (State's Historical Publications), a series of historical editions that currently comprises almost 400 volumes [2].

Of course the historical method has not remained unaltered since the nineteenth century. In his recent Historiography in the Twentieth Century. From Scientific Objectivity to the Postmodern Challenge, George G. Iggers has discussed a number of important historical approaches from this century [3]. Subsequently, his book features, among others, the French Annales historians with a fascination in the 1960s for quantifying data and the American 'New Economic History' that looked in particular to (economic) theory as a support for the analysis of history. Also 'micro history', that tries to bridge the gap between the study of seemingly unimportant details with the wider social changes, is examined, and finaly the most recent phenomenon, the 'linguistic turn', the adherents to which claim that every interpretation of the past is a construction based on texts which are as such already subjective. As Iggers explains, in the practice of historical research all these approaches are alike in that they cling to the critical use of sources. While it is true that the nineteenth century view that the historian through his research gets a grip onto the truth is no longer held to be valid, nevertheless, according to Iggers, there is an element of continuity in many more cautious viewpoints on historical knowledge: 'The historian is still bound by his or her sources, and the critical apparatus with which he or she approaches them remains in may ways the same' [4]. In other words, the constant questions for the historian are: What was the purpose of the creation of this source? Who created it? When and in what context was the source produced? What other sources are associated with this single source?

Research into sources and source criticism thereby represent the peculiar feature of 'the historian's craft' [5]. Documentary editions promise to simplify or facilitate a great deal of historical research. The demands of the source criticism have led those involved in historical editing developing a number of principles to keep in line with these requirements. These are discussed in text-books and journals and where the application is concerned they are explained in the foreword of the separate editions [6]. The first principle is one of 'comprehensiveness', the aim to process all documents in some way or another, of course within the confines of a topic or set of sources. Consequently, the objectivity of the selection criteria for documents is a principle. The editor does not select what seems to be of 'historical importance', but rather lays out objective, often formal, boundaries in order to avoid the selection becoming a mirror image of only a transient historical trend. Furthermore, documents ought to be presented in an accurate and reliable way, which has resulted in rules being set down for the transcription of texts. Another principle focuses on the context of a document, which expresses itself in an annotation to clarify a historical situation or a reference to related documents. Finally, documents should be made accessible, for example by means of an index.

In practice, these principles are flexible and can be applied in different ways depending on the nature and aim of the source edition and the means available. By way of an example: how can such a comprehensive an overview as possible of selected documents be given within the limitations of a fixed number of volumes and scarce, financial means? The text edition is the most well known and 'classic' format for historical editing. This is time-consuming and takes up a lot of room. This is why the compilation of overviews of archive materials, making short summaries of texts rather than editing and publishing the text itself and in some cases the production of just an index to a source are alternatives. The different series of the English Calendar of State Papers, summaries of state documents from the early modern age, are examples of this [7]. In the 1950s, the idea of making microfilms of sources was hit upon and combined with the publication of a finding aid as a way of accessing the microfilm. This alternative medium was more widely used in the United States than in Europe as a complement to the printed documentary edition. The pressure, in this respect, that the National Historical Publications and Record Commission brought to bear in its role as a funding institution no doubt had a lot to do with it [8]. In all these cases, efficiency was the motivating factor and so a different method to the text edition was decided upon, because a source was too voluminous to be published in its entirety.

This was the state of affairs when the computer made its entrance into the profession of historical editing. Initially the computer was used in particular to facilitate a more efficient work process, but gradually we are seeing many examples of documentary editions where modern technological applications have been widely used. This development is obvious. Indeed the disclosure of sources in catalogues, document systems and historical editions pre-eminently lends itself to the application of computertechnology. In Europe as well as in the United States the first electronic source editions have already appeared — on CD ROM as well as on the Internet — and an even greater number of projects are underway. Conceptually, the following elements can be distinguished in these editions: 1) providing information about historical documents instead of editing them together with electronic search tools; 2) applying electronic text retrieval possibilities in running text; 3) integrating images of the original documents in a documentary edition. The question that springs to mind in light of this development is: to what extent are the afore-mentioned principles, the fruit of more than one century of expertise and experience, still relevant in the new context of electronic editions of historical sources? And what are the advantages of editing a source electronically? With the help of two examples, borrowed from the practice of the Institute of Netherlands History, these aspects will be discussed.

Example 1: The correspondence of William of Orange 1533-1584

William of Orange is one of the most important figures in Dutch and early modern European history. The American historian, John Lothrop Motley made him into a hero in the nineteenth century in his book entitled The Rise of the Dutch Republic. Born into the aristocracy (prince of Orange, count of Nassau etc.) and residing in the Netherlands that belonged to the kingdom of Spain, he led the Dutch Revolt against Philip II, the Habsburg King of Spain. The matters of contention lay in the field of government and religion, inasmuch as William of Orange defended provincial and local liberties and a tolerant religious attitude against the centralism and fanatical catholicism of the Spanish monarch. In order to protect the tiny Netherlands against the mighty Spanish kingdom, William of Orange attempted to win many foreign sovereigns over to his side including Queen Elizabeth I of England and the protestant German princes. To accomplish this he engaged in an extremely lengthy correspondence. In fact he discussed the issues concerning the Dutch Revolt not only with foreign sovereigns, but also with countless local figures of authority. This is why his correspondence is of such great importance to our understanding of the European relationships, the development of modern states, the financing of wars and the birth of a republic — the Dutch Republic — in a Europe of kings. In 1584 his life ended prematurely when he was the victim of a murder, masterminded by Spain.

From this correspondence to and from William of Orange, approximately 3,500 letters were edited and published thirty years ago and were therefore known. Because it was patently obvious that the actual number must be far greater, our institute decided to establish a new research project [9]. The prospect of a text edition in printed format containing thousands of letters, often written in illegible handwriting, did not bear thinking about. This is why an inventory was decided upon rather than a full text edition. Archival research produced a harvest of approximately 10,000 letters, collected from about 80 archives in the Netherlands, Belgium, France, Spain, Italy, Austria, Germany, Poland, Denmark and Great Britain. Key points have been made about each and every letter found: the name of the correspondent and the recipient, the date, the status of the letter (e.g. draft, original, etc.), the archive where it is stored and a transcription of the first line (so that it can be recognized during an archival search). We could have left it at that, however, to further enhance the information, it was decided to add to each letter a brief description of its contents and any additional comments with regard to the letter. The plan was to publish the list of the 10,000 letters drawn up in this manner, altogether amounting to more than 2000 printed pages. To make this information accessible to the user an index of names and subjects would be added.

In the latter half of the 1980s, to facilitate the work of the project, a database was developed in which the data could be stored in a structured format. In this way, it was possible to search, check, add, select and sort data far more quickly than would ever have been possible using a traditional card index system. Subsequently, it was decided not to publish the data in book format after all, but electronically. More or less at the same time, it was decided to add images of the original letters to the data about each letter. This would compensate somewhat for the absence of a full text edition. In view of the fact that, in light of the work of the project, a photo of each letter was made that was available on microfilm at our institute, the decision regarding the enhancement of this project could be made. A new project was started a short time ago that focuses on the scanning of the microfilms of 10,000 letters, altogether about 35,000 images (a letter usually comprises more than on shot).

The advantages of an electronic edition of the inventory of William of Orange's correspondence — in the format of a CD ROM or on the Internet — are obvious. The possibilities for searching are greatly increased, simpler and more effective than is conceivable in a printed inventory. The opportunity to select, sort and combine the data per letter electronically is actually much greater than could ever be realized using a manual index. The possibilities of clicking with the mouse from data concerning the letter to the images of the letter results in a situation in which letters scattered throughout 80 European archives can be studied from anywhere in the world. The digitization of the material and the provision of data and images electronically make it possible to decentralize locally controlled archive material, so that the research can be carried out in direct relation with the data. The potential to manipulate digital images makes that the quality of the images can be considerably improved. Moreover, reading these images whether from the screen or from a print-out, is considerably easier than consulting microfilms. As long as it remains impossible to recognize handwriting styles through automated means and to convert them into digital units — and it does not seem likely that anything other than the most rudimentary progress will be made in this area in the near future — then the addition of digital images of source material to a documentary edition is a major step forward and one which we can expect to see applied more and more frequently. [10]

The scholarly principles for editing documents have been completely adhered to with this edition, even though this inventory is not a text edition in the true sense of the word. On the basis of exhaustive archival research, it can be stated that the inventory is 'comprehensive', even though the last letter will probably never be found. The key points listed for each letter on the date, individuals, etc. provide an objective entry point to the collection of letters, although as a result of only compiling key points there is an inevitable loss of information. The brief description of the contents of each letter and the added image of the letter are intended to compensate for this shortfall. Summary notes about individuals or references to letters directly linked to each other, should give the user a basic aid to place the letter in its correct context.

It is precisely this method — the extraction of elements of information from documents — that enables large numbers of sources to be made effectively accessible. This is why it may be expected that many documentary editions, especially those involved in processing vast amounts of material, will go over to this method. This statement can be illustrated with an example from twentieth century history. In the circles of editors of documents on national diplomatic history, such as the FRUS (the documents about Foreign Relations of the United States), the possibilities afforded by electronic publishing are regularly discussed. The edition of the Swiss Diplomatic Documents is currently at the top of the list. This publication has appeared up to now in printed format, but a concurrent database has been developed that can be consulted via the Internet. [11] For each document, the names of individuals, organizations, geographical elements and data are given, as are keywords, on every document. Images of the documents will be added in the future. The Swiss have not yet decided whether or not to replace the printed publication with this electronic option. They are leaning in this direction because the electronic edition would be less expensive. And what methodological objections could be raised against this?

Example 2: the Resolutions of the Dutch States General 1626-1651

The Dutch Republic, which has been mentioned above, was a federal state consisting of seven highly sovereign states or provinces. The whole country was united by a few general governing bodies, of which the States General was the most important. All seven states were represented in this body. The authority of the States General lay in the area of executing foreign politics, deciding on matters of war and peace, defending the Dutch Republic at sea and on land, and monitoring the Dutch East Indies and West Indies Companies. New Netherland, the area along the Hudson River and the Delaware River, that was under Dutch control until 1664 and afterwards for a short time until 1674 when it fell definitively in British hands, was formally governed by this States General. Access to their resolutions is only available from 1651 onwards by means of a contemporary index in the first years and later publication of the resolutions. Up until 1651 there is no way to access the contents of this important historical source.

An attempt to solve this problem was made by means of an edition. [12] Between 1915 and 1994 our institute published a total of 21 volumes with summaries of the resolutions concerning the years 1579 up to and including 1624. Furthermore, references and annotations were made in this edition. But it would have been unattainable and irresponsible to continue with this policy without adapting it. The years 1625 up to and including 1651 comprise approximately 42,000 pages of text and would require approximately 60 years' work of an editor: an unrealistic perspective considering the means available. Applying strict selection criteria to reduce the size was not a serious option: this would have adversely affected the principle of 'comprehensiveness'. This is why three options were formulated each one of which was intended, with the help of information technology, to provide access to this gigantic source: 1) the compilation of an index of names of individuals, places and subjects that are mentioned in the resolutions linked to images of the original source together with an electronic search option; 2) the compilation of summaries of the resolutions in modern Dutch. The summaries should be searchable with the aid of a text-retrieval programme; 3) to make a transcription of the seventeenth century text. The transcription should be made searchable with the help of a text-retrieval programme.

I would like to examine the advantages and disadvantages of the third option here in greater detail.

There is enough literature and plenty of examples concerning the use of text-retrieval programmes in historical research, especially in the area of content analysis. [13] Researchers have resorted to the computer for assistance in tracking down the frequency of words, their meaning and associations, and even features of style in texts. A well known American example concerns research into the authorship of a couple of the Federalist Papers. Less attention has been given to the disclosure of texts with the help of text-retrieval programmes. Can searching for words or combinations of words within a text with the help of search keys, such as the well known Boolean operators and searching for the proximity of words replace the most prominent way of accessing a text, namely the index?

The possibilities and problems have been mapped out by us in some experimental trials. [14] From the literature, familiar problems such as variations in spelling in older historical texts and the considerable noise that can be created by a 'propensity' of hits do not cause the greatest headaches. Fuzzy searching and wild cards can go a long way to solving the problem of the different spelling variants in older texts. Naturally, this technique presupposes a considerable knowledge of the text and word usage of the era from which the source stems. To give a simple example: all sorts of variants of the place-name 'Amsterdam' occur in the seventeenth century, such as Amsterdam, Amstelredamme en Aemsterdam. By using wildcards, in this instance by typing in A*mst*, it would appear that all the variants could be retrieved. It is inevitable that search commands will lead to unintentional hits because identical words can have different meanings. The context has to clarify which hits were intended. The incidence of too many hits should perhaps be viewed and accepted as a side-effect of this new technique. A much greater and deep-rooted problem is that a text-retrieval programme searches for words within the text and the historian is more interested in subjects, in abstract phenomena that are not in fact expressed as such in the text. The historian sets great store, for example, on 'social welfare' but the text only knows words such as begging, collections, poor house, and orphans. Experience teaches us that the more specific the subject matter in a given text is, the fewer the synonyms that appear in the text; and the more abstract the use of words is, all the more successful the application of a text-retrieval system will be as an alternative for the time-consuming work involved in compiling an index. A philosophical or official text lends itself more easily to a search using a text-retrieval programme than a diary or a medieval chronicle, both of which are characterized by a whimsical use of language. It has been noticed too that recent historical texts are more straightforward and abstract in their use of words and concepts than older texts. In this way, the application of text retrieval is more successful when used on texts dating from the late eighteenth century than on early seventeenth century texts. In all cases, however, a prior knowledge of the text, terms and word usage is a must. A researcher has to be well-acquainted with a source in order to perform a search with the help of the text-retrieval facility.

Should the editor of a text not wish to demand too high a level of familiarity with the historical source on the part of the user, further assistance can be offered. First of all, a series of search options can be prepared, by means of several examples or as a systematic processing of the text. In this way, it could be indicated beforehand how to search in the resolutions of the seventeenth century States General for passages on, for example, how the army was financed. The more systematic the help that is offered to the user, for example by linking key words to concrete words in the text, the closer the resemblance to a conventional index, and the less time is gained in developing a source edition. A second solution is to attach a list of synonyms and preferably a thesaurus, in which the relationships between words from a certain category are set out, such as the relationship that could exist between the example mentioned above under 'poor house' and 'social welfare'. The problem is in fact that such lists of synonyms and thesauri exist for modern, but not for older forms of the language, and in any case not for Dutch. The editor would therefore have to create it himself. Such a task would be extremely useful and would be of great importance for everyone researching older texts, but it would also be a difficult, extensive and perhaps even impossible task.

Each one of the three possibilities offers, in its own right, a method to reduce the amount of work involved in disclosing 42,000 pages of text of the seventeenth century States General. Each option also gives an added value measured against criteria such as the efficiency of the work process and the effectiveness for the user. The first option — just an index together with an electronic search tool — offers access, but creates a great distance with the source, even if images of the source were added. The subjectivity of the individual preparing the index, no matter how limited this might be thanks to, for example, guidelines, will never be completely satisfy the user who will therefore often wish to consult the source independently. The compilation of an index for this immense source is moreover, time-consuming and perhaps not so motivating for the persons who have to carry out the task and is therefore perhaps not the most preferable choice. The second option, the compilation of summaries with the aim of being able to consult them with the help of a text-retrieval programme, provides considerably more information for the user than the first-mentioned choice. Here too, subjectivity plays a role. The editor of the summaries can strive to maintain a substantial standardization for the use of words and meanings to make the text-retrieval programme as easy as possible to use. The third possibility brings the user closest to the source, but, as has been explained in great detail above, the use of text-retrieval on transcriptions of seventeenth-century texts is by no means straightforward. One could suppose that this third option would without any doubt be the most time-saving and therefore the most efficient one (transcripts can be made by less trained and less expensive personnel while the editor would supervise the whole process), but the least effective for the user and could saddle him or her with a lot of work and bother. But in each of the three options, it should be stressed, great care is taken to adhere as far as possible to the scholarly principles unique to historical editing. The text is disclosed in its entirety so that no refuge is sought in a constantly subjective selection within the historical source in order to save time. Furthermore, in every option an attempt is made to develop general access to the source.

For the time being, and possibly as an in-between step, the second variant has been selected by us, because this one guarantees better and simpler access than the more complex third variant. With respect to this third option more experiments are necessary to determine how the researcher can best be aided to conduct succesful subject searches, without the editor having to develop a complete subject index since this would be an impossible task with a source of this magnitude. Nevertheless, the latter could well become an important tool in the future in offering the user access to older texts of an extremely bulky nature. Good examples of what is possible can be found in the British Hartlib Papers and the American Papers of George Washington. The first project comprises among others the transcription of 20,000 pages of seventeenth century text in the field of intellectual, political and religious history. Search commands have been developed in advance for this project through which several subjects are disclosed. The user can also formulate his or her own search commands on the basis of words from the text. Together with images of the documents, this edition has been published on CD ROM. [15] Of gigantic proportions, but less complex is the access that the Library of Congress offers via the Internet to its own collection of 65,000 documents from and about George Washington. The user receives a digital overview of the material available, is able to consult transcriptions from part of the documents and to search using words from the text and likewise view part of the digital images of the original documents. No search commands according to subject have been developed. [16] This difference in the level of disclosure between the Hartlib Papers and The Papers of George Washington can perhaps be seen as a confirmation of what has been stated above: the further back in time one goes, the greater the need becomes for a more substantial disclosure of historical texts.

Conclusion

Let me draw some conclusions. Historical editing has a long history and has been practised professionnaly since the nineteenth century. Research of sources and critical use of sources remain characteristics of the profession even in the considerably changed practice of history in the twentieth century. Documentary editions offer a helping hand to the researcher and the editors of documents have developed rules to satisfy the demands of source criticism. The application of computer technology offers, in this respect, excellent opportunities. Particular attention has been paid above to the disclosure of massive historical sources counting tens of thousands of pages of archival material. Three methods have been discussed: providing information about historical documents instead of editing them together with electronic search tools; applying electronic text-retrieval possibilities in running text; integrating images of the original documents in a source edition. It may be clear that these new approaches bring about great changes in historical editing. But these changes do not primarily involve the scholarly principles that form the very foundation of the editorial practice. In the examples mentioned, these principles still form the basis for an edition.

It is remarkable that major initiatives in the field of electronic historical editions — such as the Model Editions Partnership in the United States — in fact have as their goal to maintain general editorial principles in a new electronic environment. [17] In an illustrative prospectus on this cooperative link between seven American documentary editions, which include the The Papers of Henry Laurens and the Lincoln Papers, different combinations are outlined of texts, scholarly commentaries such as annotations and images of documents as models for future source editions. The authors elaborate in particular on the opportunity to enrich and expand documentary editions, by for example using the possibility of offering different versions of a text — forms of transcribing; variants in the text — simultanuously, whereby the user is left with the choice of deciding what he or she wishes to appear on the screen. Also the possibility to sort documents according to the researcher's request — chronological or thematic — and to provide access to different types of annotated levels has been advocated. In this approach it is the refinement of the scholarly apparatus which enhances the possibilities to take advantage of the needs of different target-groups that is seen as the great gain in the electronic editing of historical sources.

In this paper we have paid particular attention to the advantages that the application of computer technology can have for the disclosure of immense quantities of historical sources. In the past different alternatives were developed for the text edition as the most well-known, as well as the most laborious, format for making historical sources accessible. We have described above the method of 'calendaring' texts and the compilation of finding aids to access microfilms. The three modern technological applications discussed offer ongoing positive possibilities. The nature of the source, the target group for a publication and the financial means available should determine which method is selected. To keep with the examples mentioned: in the case of the correspondence of William of Orange, that involves 10,000 letters written in different languages, varying in contents and paleographically difficult, the method of transcription and text-retrieval was ruled out because of the excessive work involved. The users of the William of Orange-database will be highly specialised researchers who mostly will be able to read the original letters. Therefore it is justified in this case to give information about the documents and to integrate the images of the sixteenth century letters in the edition. In the case of the States General however it is somewhat futile to limit access to the resolutions to providing every resolution with core data such as date, place of origin, and a global indication of the contents. The resolutions are too similar for this and this is precisely why a more detailed way of accessing the source is necessary. The general character of this source, important for any researcher who aims to make serious study of the Dutch Golden Age in the seventeenth century, requires a simple method of access. Summaries in today's Dutch will be more fitting to this purpose than transcriptions of the original text.

To summarize this paper in only a few words: the suggestion made here is that electronic historical editions can be developed within the current scholarly principles and standards, can enhance the effectiveness of the disclosure of vast quantities of source material and perhaps even increase the efficiency of the editorial process.

Notes

1. D. Knowles,Great historical enterprises. Transactions of the Royal Historical Society, 5th Series, 10 (1958-1960).

2. For a concise survey of source editions in Europe and the United States in the 19th century, see H.E. Barnes, A history of historical writing (New York 1962; second rev. ed.), ch. IX. Information on the Institute of Netherlands History, http://www.kb.nl/ing. A short introduction by D. Haks, 'The Institute of Netherlands History': A.J. Veenendaal, J. Roelevink eds., Unlocking government archives of the early modern period (The Hague 1994) 11-17.

3. G.G. Iggers, Historiography in the twentieth century. From scientific objectivity to the postmodern challenge (Hannover, Londen 1997).

4. Iggers, Historiography in the twentieth century, 144.

5. To mention only one example from an extensive literature on historical methods: M. Bloch, The historian's craft (New York 1953; originally French: Apologie pour l'Histoire).

6. For instance M-J. Kline, A Guide to Documentary Editing (Baltimore 1987, second rev. ed. 1988). Specialized journals: Documentary editing, published by the Association for Documentary Editing; Editio. International yearbook of scholarly editing.

7. B.S. Jackson, '"The Calendars of State Papers" in context', in: Veenendaal, Roelevink, Unlocking government archives, 26-41.

8. Kline, Documentary editing, 8-10.

9. There is a short outline of the project available in Dutch: B.A. Vermaseren, 'Rapport betreffende een uitgave van de correspondentie van prins Willem I van Oranje': Bibliografie van dr. B.A. Vermaseren, Th.S.H. Bos en J.G. Smit ('s-Gravenhage 1977) 25-31.

10. E.L. Helsper, L.R. Schomaker, H.-L. Teulings, 'Tools for the recognition of handwritten historical documents': History and Computing, 5, nr. 2 (1993) 88-93.

11. M. Perrenoud, 'Outils informatiques et documents diplomatiques: l'exemple d'une base de données sur les relations internationales et la politique étrangère de la Suisse': Diplomatic sources and international crises. Proceedings of the 4th Conference of editors of diplomatic documents, L. Nuti ed. (Rome 1998) 51-68; for the database, http://www.netcetera.ch/dodis.

12. J. Roelevink, 'Navigating new waters: the project "Resolutions of the States General 1576-1670"' and Th.H.P.M. Thomassen, 'The history of the archives of the States General and its consequences for their cataloguing and editing': Veenendaal, Roelevink, Unlocking government archives, 67-83, 84-94.

13. D.I. Greenstein, A historian's guide to computing (Oxford 1994) 175-199. On text-retrieval and the disclosure of historical sources, H. Voorbij, 'Analyse en ontsluiting van teksten met behulp van de computer': O. Boonstra, L. Breure, P. Doorn eds. Historische informatiekunde. Inleiding tot het gebruik van de computer bij historische studies (Hilversum 1990) 130-183, esp. 165-183; J. Roelevink, 'Waden door woorden. Inhoudelijke ontsluiting van elektronische tekst': soon to appear in VGI Cahier, 11.

14. Report Textretrieval als middel ter ontsluiting van historische teksten, http://www.kb.nl/ing/werkbest/tekstret/tekstretr.

15. Information on The Hartlib Papers Project: http://www.shef.ac.uk/uni/projects/hpp/hartlib.

16. The George Washington Papers at the Library of Congress: http://memory/loc.gov/ammem/gwhtml/gwseries.The transcriptions are taken from the printed source edition: The writings of George Washington from the original manuscript sources, 1745-1799. J.C. Fitzpatrick ed., 39 vols. (Washington 1931-1944; repr. New York 1970).

17. Model Editions Partnership. Historical Editing in the digital age: http://mep.cla.sc.edu.

Donald Haks

Institute of Netherlands History, The Hague, The Netherlands [email protected]