Dr. Judith A. Weller

There has been considerable misunderstanding by modern historians about traction by equids during Antiquity. Much of this has been fostered by Lefebvre des Noëttes' work Le cheval de selle à travers les âges: Contribution à l'histoire de l'esclavage.1 Many of the book's conclusions were widely disseminated, especially in the United States, by individuals such as T. Lynn White in his book Medieval Technology and Social Change2 who ignored the vast amount of literature which had been rapidly accumulating raising serious questions about the Commandant and his work. Consequently many Medievalists compounded the error by assuming that the horse collar was a Medieval "invention" which enabled the horse to pull heavy loads far exceeding that which could be pulled in ancient times. Now, however, we know that all this has no basis in fact. Yet even to this day books appear such as Gies and Gies' Cathedral, Forge and Waterwheel3 which do little more than compile and regurgitate in an entertaining way ideas and theories, now proven untrue, about the Middle Ages.


Commandant Richard Lefebvre des Noëttes was an Officer at Alençon from 1893 to 1901. He frequented antiquarian book shops, studied monuments and sent some articles to the Bulletin des amis Ornais. In 1896 while a captain he attempted some research into horse shoes. Transferred to Saint-Germain-en-Laye he suffered an accident which interrupted his military career. Now free to pursue his research, he studied for 2 years at L'École des Chartes under Ch. de Lasteyries and Lefevre Pontalis. He then collaborated on the Revue d'Art ancien et moderne, participated in the works of the Société national des Antiquaires de France.He continued his interest in the question of horseshoes and in 1900 he conducted an experiment on the Roman Hipposandal.4

In 1910 he conducted his experiments in Classical traction. While conducting these experiments he was studying iconography and he based his design of the harness system on his study of available iconography5. There is no indication in his publications that he ever sought assistance or expert advice in reconstructing his version of Roman harness6. If he received such technical assistance he did not mention it.

This experiment apparently did not attract any attention at the time and no mention can be found of it in scholarly contemporaneous articles. In any case World War I and its aftermath put Lefebvre des Noëttes' work on hold. However, in 1924 he published the results of his studies and experiment in a book entitled La Force Motrice à travers les Âges7. He purported to study all forms of animal traction in both the ancient and modern world. His book is full of discussion of Ancient Egypt, the Han period in China, harness systems of bovids and equids8. He maintained that his studies demonstrated that the harness systems of the ancient world were defective and that they caused pressure on the horses' windpipe causing the animals to be strangled by their harness if they advanced too quickly or pulled too heavy a load. He maintained that the ancient harness system enabled the horse to pull only light loads9. Even after the 1924 publication there is no evidence that the book found a widespread support.

So why did the 1924 version lapse into oblivion but yet the 1931 edition form the basis for a fundamental misunderstanding of early harness system? An important clue is to be found in the Preface of the 1931 edition. The 1924 edition contains no preface, merely an introduction by the author. But the 1931 edition contains a laudatory preface by none other than Jérôme Carcopino, the leading French classical scholar of the period10. More interestingly the book was expanded and had a new title. The new title emphasized only the horse and had as its subtitle: "Contribution to the History of Slavery." This new edition now emphasized a conjunction between horse harness and slavery in the Ancient World. The basic thesis was that the because of the defects in both Roman and Ancient Harnessing systems, slaves were used to replace animals for ploughing, hauling heavy wagons. etc."

This sociological turn had been mentioned in the 1924 edition11,but was not the principal emphasis of the work which methodically set out to demonstrate the basic similarity of all harness systems in the Ancient World. In writing the preface to the 1931 edition Carcopino used his "scientific authority" to place a seal of approval on this work and its thesis12. Carcopino heaped profuse and fawning praise on Lefebvre des Noëttes for his marriage of philology, archaeology, and epigraphy13.

However, what interested Carcopino most about the work, was the nexus which Lefebvre des Noëttes made between slavery and the weakness of ancient traction systems14. Nor was Carcopino the only person to heap effusive praise on the Commandant15 but by 1938 the Commandant's work was under attack. and his supporters were publishing defenses of the book16.

Despite the attraction Carcopino and his circle had for this sociological connection, it was the first part of the work to fall to a more scholarly analysis in the post-war era. It took until 1957 to demonstrate17 that Lefebvre des Noëttes was in fact basing his argument on incomplete evidence and that his theory had been accepted simply because historians and archaeologists were ignorant of mechanical science18. Today this theory of a connection between slavery and harness which Lefebvre des Noëttes' propounded in the 1931 edition has no supporters.

What is amazing, however, is that this officer's view of harness systems in the Ancient World persisted with little or no critical evaluation for almost half a century. Men who were considered scholars in the field accepted this work without ever asking themselves if there were any flaws in the methodology or the conclusions. More importantly, did this obscure Commandant have the qualifications necessary to interpret correctly the iconography that he was using as support for his work? As the result of later experiments and scholarly work, we know that he did not. But we still need to understand why the errors in his work have endured for so long. There is no doubt that World War II and the chaos of the early post war years ended the opportunity to analyze thoughtfully his work. The 1960's and 1970's were the years in which his ideas became entrenched, more in the mind of Medievalists such as White than in the minds of classicists who were the first to see the flaws in his methodology. So the demolitions of Lefebvre des Noëttes' theories about Roman Harness are to be found in the writings of classicists and archaeologists who, equipped with new tools and new discoveries, demonstrated the fallacy of the 1924/1931 work.


For the purposes of this article, I am using the following dates to define various historical periods. The focus will be on western Europe as that is the area where the horse collar eventually became widespread in the Middle Ages and is the area most written about in any discussion of traction systems.

End of the Western Roman Empire:
476 AD (death of Romulus Augustulus) - although a more practical date might be considered to be 410 AD when Alaric sacked Rome.
Dark Ages:
410 AD - 800 AD (Coronation of Charlemagne.)19
Middle Ages:
800 - 1500 AD. The period after 1500 is considered to be the Renaissance although the Renaissance has, in reality, a variety of dates ascribed to it depending upon the geographical area of Europe of which one is speaking.

  1. Lefebvre des Noëttes. Le cheval de selle à travers les âges: Contribution à l'histoire de l'esclavage.Paris, 1931
  2. White, pp. 59-62
  3. Gies and Gies , pp. 45-47 - see p. 46 : "The new device replaced the old throat and girth harness which choked the horse as he pulled against it."
  4. He included information about this experiment and a photograph of his reconstruction in his 1924 worked. Lefebvre des Noëttes, 1924, pp. 61-65, photograph fig. 92
  5. Amouretti, pp. 219-220)
  6. Lefebvre des Noëttes. 1924, pp.85-86
  7. Lefebvre des Noëttes, Paris, 1924
  8. Lefebvre des Noëttes, 1924, passim
  9. Lefebvre des Noëttes, 1924, pp. 85-86. Let me also take this opportunity to silence those critics who would cry out that only the 1931 edition should be used. Not one word has been changed in the section on "Les expériences practiques" between the 1931 edition on pp. 162-164 and the 1924 edition of pp. 85-86. Actually the most noticeable difference is that photographs of the experiment are to be found in the 1924 editions, whereas only crude tracings of those same photographs appear in the 1931 version. Also the 1931 version is in 2 volumes with the second volume being devoted to photographs of statues and iconography on the topic.
  10. Lefebvre des Noëttes, 1931, Preface, pp. i-vii.
  11. Lefebvre des Noëttes, 1924, pp. 30-31
  12. Raepsaet, 1982, p. 215.
  13. Lefebvre des Noëttes, 1931, p.v
  14. Ibid., pp. i-vii
  15. J. Moulinier,
  16. Amouretti, p. 22 see also fn. 8, & 9
  17. Salama, p. 291
  18. Chevallier, p. 179
  19. I have no wish to get into a discussion of the term Dark Ages. I find the term convenient. I view the period between 410-800 AD as a sort of Dark Ages which comprises the idea of a collapse of central government, the gradual weakening of Roman cultural, linguistic, legal and political institutions. It was period of great chaotic change caused by the inroads of invading tribes out of the east - it is also a period for which we have little written, or even cultural evidence. These sorts of evidence may have existed but been destroyed in ensuing military actions. Monasteries made some effort to preserve culture and on some of the manuscript which have survived we have crude pictorial evidence of activity such as the Trier Apocalypse. Nevertheless it is not a period about which we have nearly the amount of knowledge of the society which we have for the period preceding it (the Roman Empire) and after it (Charlemagne). Consequently we have no continuous, clear path which enables us to document with a high degree of certainty the changes in the life of the society.