Proceedings 32        22 April 1993        St Peter’s College Oxford

Two related papers were given:

Norman Smith on “Sources of Medieval Technology with special attention to Byzantine and Islamic Contributions”.


Dennis Simms on: “Was There a Technological decline and Fall of the Roman Empire?”  (see Proceedings No 33)



The summary which follows is an article by Dr Smith himself based on his talk that day and with a slightly different title.



Such has been the preoccupation with the technologies of the Roman empire and the medieval world that the period between is rarely considered. This is a pity because some of the more interesting aspects of technology's history are its continuity from one culture to another and the nature of the processes by which ideas and methods are diffused and transmitted. And in any case, for the Middle Ages the question of origins is an interesting one in itself.

In principle, at least six inputs to the medieval technological scene can be nominated: a Roman legacy in practical technology; the discovery of Roman texts; an Islamic influence via Spain or Sicily; contact with the East through the Crusades; transmissions from the North, Celtic or Germanic; and of course, developments originating in the Middle Ages themselves. In what follows are summarized some of the sorts of instances which arise with an indication of the difficulties of turning fragmentary evidence into firm conclusions.

The Piston Pump

Dr G.J.Hollister Short is quite correct when he nominates the piston pump as one of the more mysterious of early machines although one is bound to observe that an inadequate terminology has not aided the enquiry. How piston pumps work is not completely clear to everyone who writes about them, and the evidence is scarcely ever marshalled in its entirety. The Romans appear to have utilized the piston pump frequently. Its archaeological evidence from mines, wells and ships suggests application to firesetting, water raising and bilge pumping and Hero of Alexandria prescribes it for fire-fighting. My own opinion is that even at an early date the piston pump was, at least partially, a suction pump and, however imperfectly the principle was understood, its exploitation made for a much more practical setting of the piston pump on ships. And as we know from al-Jazari, his own suction force pump for irrigation was 'like the ejectors of naphtha except that it is larger.’ In other words we can suppose that Byzantine navies used suction force pumps to project Greek fire. Given the ingredients of the mixture used, the submerged pump by comparison, does seem hopelessly unworkable.

So, we seem to have a continuous history of piston-pumping in and around the Mediterranean for 1500 years. When at last the suction lift pump appears in a late medieval European source, namely Taccola’s De Ingeneis most likely we are looking at a Roman legacy which has either been preserved within Europe or has returned via Byzantium and Islam. Taccola’s pump is distinguished by two particular features'- The pump barrel is, essentially, a tube continuous from sump to outlet, while the piston is itself valved. Such features seem to me to be the result of adapting the piston pump to work in mines. Very possibly Taccola’s pump came to northern Italy from southern Germany.


The origins of water-wheels may not be quite as obscure as piston pumps but it is a markedly more contentious subject with recent work having produced as much confusion as clarification. It is sufficient here to summarize a few points. The ancient world knew and used the vertical water wheel and an installation such as Barbegal reflects a confidence and commitment which is significant. Whether or not the Romans knew the horizontal water-wheel as well is a point I have touched on elsewhere.* A key issue here is whether we are looking at the invention of a concept or the invention of a machine. Some writers evidently imagine that water-power was developed as an idea from which a range of alternative designs of mill sprang readily off the drawing board, so to speak. However, it may be that an existing mechanical tradition evolved a machine first and that the concept came later Something of the same sort of problem is an issue in the origins of the windmill.

What we can be sure of is that Roman water-wheels were applied exclusively to corn-grinding and this technology was inherited by the Middle Ages. The quality of an invention however is not to be confused with the quantity of its application. The significance of water-power in Roman times is that it was exploited at all: in the Middle Ages what is notable is that water-power began to be applied to a variety of new and different tasks, such as sawing, fulling, paper-making and iron working. The new wisdom, contrary to Lynn White Jnr., is that this widening of applications did not revolutionize the medieval economy. On the other hand technological changes were imperative and the machinery being driven by water-wheels and windmills displays considerable and ingenious development, full of significance for the future.


In Medieval Technology and Social Change, and other of his writings, Lynn White Jnr. was, on the whole, disinclined to consider medieval technology.’ classical origins. When he tried, the results were not always either accurate or convincing. His

*see SHMTS Proceedings 22, .Barbegal and the Roman legacy of Water Power. February 1991

dismissal of the evidence from Maiden Castle f Roman nailed horseshoes on the grounds that rabbits had hauled the items downwards from later levels is no more than comical and in any case is at odds with evidence from elsewhere. Moreover, Lynn White's insistence on the appearance in the Middle Ages of the heavy horse should really read heavier horse and then only latterly. In fact the earliest medieval depictions of horses do not seem to show any significant change from Antiquity in either the size of the animals or the way they are harnessed. Only at a relatively late date does the horse appear to become sufficiently heavy and powerful to even begin to compete with the ox and thereby make worthwhile the horse collar which Lynn White rated as so significant. How far this improvement of the breed was brought about by the injections of new blood from Arab animals is debatable. **


Traditions in building are probably the best known, and the most certainly attested, elements of the Roman engineering legacy to the Middle Ages. In the Roman world large-span buildings were of two types: the long basilica, such as that of Maxentius, or the domes rotunda, for example the Pantheon. The characteristic structural features were semi-circular arches, hemispherical roofs and groined vaults. The steady evolution of these devices is fundamental to the development of Gothic building, exemplified above all by the cathedrals.

The Gothic cathedral might be said to "comprise four characteristic features: [1] a load carrying masonry structure using wa11s, panels and windows to enclose great space [2] pointed arches [3] rib-and-panel vaulting and [4] buttresses to balance loads, i.e. weight. The first of these is the achievement of Gothic building; but the other three each have antecedents. In Islamic building, for example, the pointed arch began its evolution as an alternative to Roman semi-circular form. Also and Islamic building some anticipation of rib-and-panel vaulting is detectable however limited it was to decoration. Buttressing in a crude form is commonplace throughout early building while the flying buttress appears to be a later refinement. Islam's role in connecting Roman origins and Medieval achievements is not confined to structural design as such. Architects and workmen were also influentlal by demonstrating methods and improving workmanship; sometimes they were active quite far north in Europe.


Finally an example of the written legacy, the problem of Vitruvius. There are three separate questions, by no means alwaysproperly differentiated: was Vitruvius's MS known in the Middle Ages? Was he used as a technical manual? And was he influential in some fundamental sense. That Vitruvius was known is clear enough because some eighty copies of his manuscript dating from about AD 800 and later are now known. But the use of Vitruvius as a source of information is more problematical. References to the design of buildings by rules of geometry and proportion are not very conclusive because such methods were well established and routine, not only in the Roman world and then the Middle Ages but much later on as well. There is such a thing as standard practice and it has to be recognized that the instructions of Vitruvius, Villard de Honnecourt and others may appear derivative simply because they are describing techniques and devices which did not change over the centuries.

On the other hand there is some evidence rather less circumstantial. The majority of Vitruvian manuscripts descend from a single original probably written in Charlemagne's palace scriptorium about AD 800. Around AD 805 Alcuin (#) mentions Vitruvius to Charlemagne in a letter and possibly he had himself brought a copy from England. Vitruvius is also mentioned in a letter by Charlemagne's biographer, Einhard. This 9th century evidence I find persuasive because, firstly, of Einhard's involvement in building and secondly the early date. As Carolingian Europe attempted to re-establish monumental building one can see how Vitruvius might well have been a useful guide. Later on he became dispensable. It is difficult to believe that Romanesque builders had much to learn from Vitruvius and Gothic ones even less. Subsequently, in the Renaissance, Vitruvius's elevated status is an entirely different matter.


The Roman engineering legacy and the way it was passed on is a complex subject and not at all well studied. It is interesting to note that in the original The Legacy of Rome, edited by Bailey as long ago as 1923, there were chapters on Communication, Architecture, Building, Engineering and Agriculture. By comparison in the recent version, edited by Jenkyns in 1992, only the chapter on Architecture is retained.

[#] Here, as with the bulk of Europe's Latin literary heritage, we are indebted to the work of churchmen. Also, in this age of Euro-culture it may be of interest for members to know that in a recent TV costume drama on the reign of Charlemagne broadcast on French TV and made by a German-French based consortium Charlemagne is depicted as a French monarch (his troops even brandishing banners of the fleur de lys [!]. And Alcuin of York (played by Frank Findlay) makes two brief appearances as a kind of magician. (GSH)

see Proceedings 15, 'Horsebreeding in the Middle Ages,' R.H.C. Davis, October 1989 (ill.'; compare: SHMTS Proceedings 5, 'Animal Powered Machinery--The Medieval Period', Kenneth Major November 1987 (ill).