Was Roman Science in Decline? (Excerpt from My New Book)

My new book The Scientist in the Early Roman Empire will be in print in just a few days! You can already pre-order it in print or kindle. But in celebration and promotion, I’m here producing an excerpt from it. You can read a bit more about the book at Pitchstone. Next week I begin recording the audiobook. This brings to completion the publication of my Columbia University dissertation, expanded and revised for a broader audience, but still thoroughly academic (it’s heavily footnoted; the bibliography alone extends over sixty pages). The basic table of contents reads:

  1. Introduction
  2. The Natural Philosopher as Ancient Scientist
  3. The Roman Idea of Scientific Progress
  4. In Praise of the Scientist
  5. Christian Rejection of the Scientist
  6. Conclusion

But packed in there is a lot more, including full surveys of the history of ancient science and technology, and such sections as “The Scientist as Hero in the Roman Era,” “The Scientist as Craftsman in the Roman Era,” and “The Methods of Roman Scientists.” Probably the most thorough treatment of Roman science in over forty years. Here is one such section, 3.7, “Was Roman Science in Decline?” Which I’ve also added hyperlinks to for this presentation. It starts just after I finished proving with extensive evidence and citations of scholarship that Romans were fully aware of and praised past and continued scientific progress, and were still engaged in producing it. This section refers to other parts of the book, but you’ll get the idea. You’ll see hints of many other gems in there to find. So you should buy a copy!


Was Roman Science in Decline?

Nevertheless, some scholars have claimed science suffered a stagnating decline during the Roman period. Lucio Russo even claims (absurdly) that “the Romans were not interested in science” and had abandoned the superior aims and methods of their Hellenistic forebears.[775] Such assertions are ultimately baseless. There is no evidence of any difference, much less decline, in scientific aims or methodology between, for example, Hipparchus and Ptolemy, or Herophilus and Galen (Russo’s favorite examples). Only by romanticizing Hellenistic scientists, and imagining (implausibly) that they never held or defended any absurd or erroneous beliefs, can Russo contrive any appearance of decline. A more frequent mistake is to compare Hellenistic scientists with Roman laypersons, as if Hellenistic laypeople would come out any better in comparison with Roman scientists. Myths of a ‘Roman decline’ are thus often based on assessments of lay authors like Pliny the Elder, rather than actual Roman scientists like Dioscorides, Hero, Marinus, Menelaus, Ptolemy, Galen, or Soranus.[776] But the mistakes and flawed or inexact methods of an author like Pliny tell us only about the standards and practices of lay admirers of science, not what actual scientists were doing.

Of course, negative assessments of Pliny’s merits are also often exaggerated..[777] But more importantly, an individual author does not always represent their whole society—one need only compare Pliny’s treatment of medical science with that of his predecessor Celsus to see how superior a treatment the same subject could receive from another lay author of the very same time.[778] Even picking on individual scientists is not always apt. Hero might not always appear as rigorous and brilliant as Archimedes, but that may be the very reason why Hero’s works were preserved, and not those of even Archimedes, much less other Roman engineers who may have been similarly rigorous and thus too unintelligible to medieval antiquarians to warrant copying. One can only wonder, after all, what happened to the mechanical writings of Menelaus, Apollodorus, and Ptolemy, much less authors unknown. In the same fashion, one cannot claim Strabo’s failings in geography or astronomy were symptomatic of the Roman era, when that same era also produced the superior work of Marinus and Ptolemy in those same fields, and especially when it cannot be established that none of Strabo’s Hellenistic predecessors were any worse than he was.[779]

Nor does it make any sense to maintain there was a “resurgence of religious enthusiasm” in the Hellenistic age that worked against scientific advancement.[780] There is no good case to be made that religiosity and superstition was ever in any state of decline. Skepticism and rationalism remained as present but as uncommon as ever, hardly more than the preoccupation of a rarefied elite, while superstition and irrationality remained the norm against which exceptional men had battled even in Classical Athens.[781] And though ancient scientists in every era had embraced bad ideas, and did not follow their own recommended methods as consistently as we would like, the very same could be said of the savants of the Scientific Revolution. Galileo’s ideas about tides and visual rays were often wildly wrong, Kepler was obsessed with the harmony of the spheres, and Newton pursued alchemy and worked profusely on biblical theories of history, prophecy, and cosmology, spending considerable time trying to predict the apocalypse.[782] Meanwhile, bloodletting continued as a ‘scientific’ medical treatment well into the 18th century. The 19th century became an infamous age of medical quackery. So we moderns are in no position to judge.

The first half of this chapter has already shown how claims of both scientific and technological stagnation under the Romans are implausible. Peter Green concedes that “progress of a sort did take place” but then claims there remained “a dead-weight legacy from the past that in many ways made true progress virtually impossible,” a judiciously meaningless statement, since he does not explain what we are supposed to count as “true” progress or why.[783] We have already seen that scholars like Peter Green are obsessed with finding fault with what the ancients did not invent or discover, while ignoring almost everything they did invent and discover, and then accusing them of having invented and discovered nothing. Which they then proceed to explain with one or another fanciful hypothesis. It is a peculiar way of doing history. As an example at the very nexus of science and technology, Green complains that the ancients failed to invent “steam gauges, thermometers, microscopes, telescopes, [and] fine-calibrated lathes,” as if these were somehow obvious and easily conceived technologies, while ignoring the countless instruments ancient scientists did invent to further their research.[784]

More absurd is Aubrey Gwynn’s claim that “the Roman Empire never produced a scientific discovery that has been of permanent use to mankind.”[785] Even a lot of obsolete science was still a necessary step toward modern science. For instance, Ptolemy’s law of refraction was not entirely correct, but it was close, and his idea and procedures for experimentally discovering a mathematical law of refraction were certainly of permanent use to mankind, and though Hero’s experimentation with steam-powered machinery did not lead immediately to a practical steam engine, it was a necessary first step that eventually inspired it, so Hero’s discovery that steam could be used to produce mechanical motion was of permanent use to mankind. Meanwhile, many Roman discoveries (such as in pharmacology) were certainly of permanent use to mankind, or may have been yet were lost, while others (like electroshock therapy) remain in use, even if in different applications. Roman discoveries still (more or less) in use include Ptolemy’s system of cartographic projection, Hero’s principle of least action in reflection, Galen’s experimental discoveries relating to kidney function, the spherical trigonometry of Menelaus, and the idea of symbolic algebra of Diophantus—we just do not use these same systems of trigonometry and algebra today any more than we speak Latin or ancient Greek. Ptolemy’s most crucial innovation in planetary theory, the acceptance of inconstant planetary velocities and proposing a law of planetary motion (equal angles in equal times), turned out to be essential to Kepler’s solution for the planetary motions and orbits (updating Ptolemy’s law to equal areas in equal times), while the efforts of Ptolemy and Galen to unify their sciences and epistemologies were of even more general benefit to modern science. And then there were useful discoveries we often ignore. For example, one of the areas Galen knew he was making considerable advances in was the physiology of voice and speech, pursuing a comprehensive research program involving extensive physiological and anatomical observations and experiments on every related organ from the lungs and thorax to the nerves and muscles of the throat, larynx, tongue, and more.[786]

Like Gwynn’s antiquated nonsense, most of the claims of a Roman decline are so contrary to the facts that they hardly need refutation. The most famous example is a raft of assertions by Samuel Sambursky, all plagued by fanciful and inaccurate conceptions of ancient science, many of which have already been exploded in previous sections of this chapter.[787] Ancient scientists were not isolated from each other, but enjoying frequent communication and interaction, and the sharing and accumulation of results.[788] There was no relevant disdain among them for shopwork and technology.[789] There was no significant opposition to changing or interfering with nature.[790] There was no aversion to experiments.[791] There was no failure to mathematize the study of nature.[792] They actually did understand natural processes mechanically rather than organically.[793] And there is no evidence of any significant ‘rise’ in irrationality under the Romans (at least before the 3rd century A.D.).[794] Everything else Sambursky proposes confuses the effects of the Scientific Revolution with its causes, and thus fails to explain anything even when true.[795]

So when Sambursky claims a fictional stagnation resulted from a “lack of systematic experimentation and the consequent stagnation of technology, and the failure to develop algebraic notation and to introduce mathematical symbols and procedures in the description and explanation of physical phenomena,” we already know every single one of these claims is false.[796] The Romans were seeing progress in all. And even Sambursky knew he had to qualify his remarks, admitting the Romans held a “greater regard for observational evidence and an increasing demand for a more accurate description” and were conducting systematic experiments that “led to conclusions which conflicted with Aristotelian conceptions about the nature of light” and other subjects. Hence, he concludes, it was really only after the era of Galen and Ptolemy that “the combined effects of the irrational tendencies within neo-Platonism and of the anti-scientific attitude of the early Church,” and the general decline of educational institutions everywhere, finally put an end to scientific research.[797] On all that, at least, he was correct.

Similarly, Ludwig Edelstein once claimed “ancient science remained relatively useless” and “changes which in principle were within reach were actually not made” because empirical scientists were too skeptical to theorize, theorists were too disinterested in empirical research, and everyone was uninterested in controlling the natural world through technology.[798] But not one of these assertions is true, as any perusal of Galen, Ptolemy, Hero, or Vitruvius would easily reveal. More credible but still dubious is Peter Green’s assertion that “quantitative methods, essential to true scientific progress, were conspicuous by their absence” among the Romans.[799] But he still never explains what he means by “true” scientific progress, or even “quantitative methods.” Was all the scientific progress I just documented ‘fake’? Was measuring doses of medications, angles of refraction, mechanical advantages, or velocities of planets not ‘quantitative’? There were certainly many failings in the way ancient science was conducted, but an absence of quantitative methods was not among them. At most one can say such methods were not more widely exploited than they could have been, but there was no evident decline in this respect.

Peter Green has voiced many other absurd allegations. For example, he claims “the enormous weight of [Aristotle’s] authority” did “more to hold up the progress of astronomy than any other single factor,” yet progress in astronomy was not held up, and as we have seen, Aristotle’s authority was not particularly great in antiquity (in fact it was greater at the dawn of the Scientific Revolution).[800] Green claims the Hellenistic trend toward moral philosophy “culminated in the abandonment of true research” and a “reversion” to excessive theorizing, but he never identifies any point in time when the research he has in mind was “abandoned” or when theorizing was not excessive. To the contrary, Presocratic science was heavy on theorizing and light on research, while most science after Aristotle leaned quite the other way, with moral philosophy and scientific advances increasing in tandem.[801] Green also claims scientific progress in antiquity was hindered by a “prejudice” against written texts, but there is no evidence of this, any more than lectures and internships indicate any such thing now.[802] Likewise, “the subordination of experimental science to philosophical system-building” was true all throughout antiquity (in fact reversing this was a defining feature of the Scientific Revolution), yet progress continued.[803] Likewise, the fact that, as Green says, logico-deductive conclusions are more reliable than empirical ones is a fact made much of even by Descartes and recognized still today. Though entirely true, this fact has had no effect on science now, nor did it then.[804]

In a similar fashion, Joseph Ben-David repeats one of Sambursky’s indefensible claims, that ancient “scientists built their individual systems without reference to those of others and established rival schools which, like so many religious sects, did not communicate with each other.”[805] Again, he is wrong on all counts. The works of Ptolemy, Hero and Galen are full of references to, adoptions from, and improvements upon the work of numerous predecessors in their respective fields, while Galen’s writings are filled with evidence of a lively public interaction among contemporary scientists.[806] There is no evidence that any ancient scientist behaved differently. And while there were many “competing schools of thought” on crucial questions of method and epistemology, these were not isolated nor even dogmatic enclaves, but loosely-affiliated groups of researchers regularly engaged in improvement, intercommunication, and debate. The most successful scientists, in fact, refused to align themselves with any one school, but instead learned and borrowed from them all, a phenomenon of ‘eclecticism’ that typified the entire intellectual atmosphere of the Roman period.[807] This is quite evident in Ptolemy, who merged the epistemologies of all the major schools into a practical proto-scientific system, and in Hero, who loved trumping sectarian dogmas with physical demonstrations, and in Galen, who railed against the very idea of distinct schools of medical thought and instead embraced elements of many different schools, criticized the rest, and synthesized a nearly modern combination of deductive and empirical methods of his own.[808] Galen also sought to unify formal logic by developing a comprehensive system from of the doctrines of several schools.[809]

Moreover, Hero, Ptolemy, and Galen all insisted upon the use and methodology of mathematics in the sciences.[810] And all employed systematic experiments in their work. In his Pneumatics, for example, Hero begins with a physical theory, describes experiments that establish its basic principles, affirms that such experiments conclusively refute all armchair philosophical arguments against the conclusions thus demonstrated, and then moves on to describe an extensive series of technological applications of the theoretical principles just demonstrated.[811]

We can see the same trends in the scientific writings of Ptolemaïs by the 1st century A.D. Though her books were not preserved, surviving quotations show her attacking those who divided her science into sectarian dogmas. She argues instead that to get to the truth one must unify the best elements of competing sectarian approaches and discard the rest. She criticizes those who rely on reason and theory and ignore or discount observations, and also those who only observe and ignore theory. She defends instead the need for a unified theoretical and observational approach to harmonics, integrating empiricism with mathematics. This is essentially what we also hear from Hero, Ptolemy, and Galen, and the generalizing nature of her remarks suggests she would have agreed with their extension of the same principles across the sciences.[812] Hence the Roman trend in ancient science was not as Ben-David claimed, but in exactly the opposite direction: toward communication, unification, and integration of the best elements of science and philosophy into an increasingly superior methodology.

So all these arguments for decline don’t hold up.

Besides those, however, there are four other arguments that appear repeatedly in the literature, which purport to prove that the ancients had no conception of scientific (and technological) progress or were even hostile to the idea. It is often claimed the ancient slave system discouraged interest in progress, or that progress was blocked due to the Romans being dead set against the idea of changing or interfering with the natural order, or that they never had the idea of explaining nature and natural processes mechanically (rather than, say, organically or supernaturally), or that they were so obsessed with a cyclical model of time that they were incapable of even imagining progress or thinking it possible or worthwhile. All false.


The following sections then cover in detail “The Slavery Thesis” (pp. 250-53), “Changing Nature” (pp. 253-58), “Mechanizing Nature” (pp. 258-63), and “The Cyclical Time Thesis” (pp. 263-69). And I follow that with a complete survey of “Ancient Tales of Decline” (pp. 270-307).


[ 775 ] Russo 2003: 266 (he offers several negative assessments of Roman science, none of which are demonstrated by any adequate evidence: 15, 215, 231-41, 264-70, 282-86, 318; yet ironically he challenges the basis of such assessments from other authors: 197-202). Contrast Russo’s assessment with that of Chevallier 1993.

[ 776 ] An example of this error on a grand scale is Stahl 1962 and 1971.

[ 777 ] For modern mis-assessments of Pliny (and their recent correction) see earlier note.

[ 778 ] Pliny discusses medicine in the 29th book of his Natural History, Celsus in his extant volumes On Medicine, both in Latin. Romans did not all agree with Pliny, e.g. Aulus Gellius (in Attic Nights 10.12) takes Pliny’s credulity to task, somewhat unfairly according to Beagon 1992: 11 n. 31, but there would have been many laypeople of the day who could correct Pliny on many points. Similarly, Quintilian (in Education in Oratory 10.1.128) complains that Seneca was a brilliant man but relied too much on research assistants who sometimes led him into error.

[ 779 ] For a relatively (and sometimes unfairly) negative assessment of Strabo see Aujac 1966.

[ 780 ] P. Green 1990: 481. Farrington 1965 attempts a similar but even more inept argument.

[ 781 ] See Dodds 1951 and note in chapter 1.3 on religious persecution in Classical Athens.

[ 782 ] On various absurdities among 17th century scientists see: Russo 2003: 355-59, 363-64, 366-69, 385-88 (likewise, for Newton, Rossi 2001: 203-29); more examples in Zimmermann 2011. Ultimately there was nothing any more boneheaded in ancient scientific treatises than can be found in even the most respected authorities of the Renaissance.

[ 783 ] P. Green 1990: 480-81. Similarly, Moses Finley concedes “there were improvements of one kind or another,” and “technical refinements,” but insists these were only “marginal” and not “radical improvements,” without defining either ‘marginal’ or ‘radical’ (Finley 1985: 109, 114).

[ 784 ] P. Green 1990: 481. He also complains of a lack of formal statistics and “advanced technical instruments” in antiquity (P. Green 1990: 457), even though neither existed until after the Scientific Revolution. Likewise for Zilsel’s complaint that they didn’t have periodicals (Zilsel 1945: 327). However, as noted in the previous section, the Romans must have had more advanced lathes than we are otherwise aware. Indeed it is ironic that (as also noted in the previous section) Green cites ancient precision tooling of nested cylinders to within a tenth of a millimeter, and yet he somehow thinks this was achieved without fine-calibrated lathes, which would have been needed for turning the wax molds to such a precise clearance.

[ 785 ] Gwynn 1926: 146. Such dismissiveness, which can still be found (we opened with an example from Russo), is rightly criticized in Nutton 2013: 13 -16 (though only for medicine, his remarks are as relevant for astronomy and physics) and also challenged by Chevallier 1993.

[ 786 ] See Galen, On the Doctrines of Hippocrates and Plato 2.4.33 -39 and On the Uses of the Parts 7.14 (= M.T. May 1968: 367).

[ 787 ] Sambursky 1962: 253-76. The same points (or claims even more ridiculous) are still echoed in more recent scholarship, e.g. Vernant 1983: 294-295 and 366, Reynolds 1983: 32-35, Lewis & Reinhold 1990: 2.210, and Stark 2003: 151-54 (with the same material almost verbatim in Stark 2005: 17-20, though adding more false assertions about ancient science and technology in 2005: 12-17). All of which are adequately refuted by the substance of the present chapter. As also (more succinctly) in Carrier 2010 (supported by Efron 2009).

[ 788 ] Contrary to Sambursky 1962: 254-55. See examples in chapters three and four here, and throughout Carrier 2016.

[ 789 ] Contrary to Sambursky 1962: 257-60. See discussion below and in chapter 4.6.

[ 790 ] Contrary to Sambursky 1962: 259-62. See discussion below.

[ 791 ] Contrary to Sambursky 1962: 261-70. See examples throughout sections 3.1 through 3.6.

[ 792 ] Contrary to Sambursky 1962: 270-73. See examples in sections 3.3 and 3.4 (and even in biology: see discussion and sources in von Staden 1996: 88-90).

[ 793 ] Contrary to Sambursky 1962: 273-74. See discussion below.

[ 794 ] Contrary to Sambursky 1962: 260-61, 274. See remarks above.

[ 795 ] Sambursky 1962: 254-56. On the relevance of this point see section 1.1.

[ 796 ] Sambursky 1963: 62.

[ 797 ] Sambursky 1963: 63-66.

[ 798 ] Edelstein 1963: 24-27.

[ 799 ] P. Green 1990: 470.

[ 800 ] P. Green 1990: 459.

[ 801 ] P. Green 1990: 482. It should also be noted that Green has been deceived by medieval selectivity in preserving texts, creating the illusion of a rising interest in moral philosophy at the expense of physics and logic that actually never happened in antiquity: see Carrier 2016: 102-04.

[ 802 ] P. Green 1990: 457. See the more reasonable analysis of Alexander 1990 and discussion in chapter seven of Carrier 2016.

[ 803 ] P. Green 1990: 481.

[ 804 ] P. Green 1990: 457.

[ 805 ] Ben-David 1991: 301.

[ 806 ] Ptolemy’s Almagest, Geography, and Harmonics are good examples of his discussion of predecessors and his reliance and improvement on them, as are Galen’s many treatises on anatomy and pharmacology, and likewise Hero’s Pneumatics.

[ 807 ] See discussion in Carrier 2016 (index, “eclecticism”) and section 3.2 above. For further discussion of the eclecticism of Galen and Ptolemy see: Gottschalk 1987: 1164-71. For Ptolemy: DSB 11.201-02 (in s.v. “Ptolemy”). For Galen: Hankinson 1992. For Hero: Tybjerg 2005: 214-15. Galen specifically describes and advocates eclecticism in On the Affections and Errors of the Soul 1.8 and 2.6-2.7 (= Kühn 5.42-43 and 5.96-103) and Seneca effectively does the same in Moral Epistles 33. See also the ‘eclectic’ credo advocated in Celsus, On Medicine pr.45-47.

[ 808 ] For Ptolemy’s scientific epistemology: Huby & Neal 1989; Long 1988: 176-207; A.M. Smith 1996: 17-18; Barker 2000. For Galen’s scientific epistemology: Frede 1981; Walzer & Frede 1985: xxxi-xxxiv; Iskandar 1988; J. Barnes 1993; Hankinson 1988: 148-50, 1991a: xxii-xxxiii and 109-10, 1991b, and 1992; M.T. May 1968: 45-64. For an early summary of both: Edelstein 1952: 602-04. Ptolemy’s On the Criterion and Galen’s On Medical Experience are prominent examples, as also Galen’s On the Sects for Beginners and An Outline of Empiricism (for all three see translations and discussion in Walzer & Frede 1985, esp. xxxi-xxxiv), as well as his synthesis of epistemologies in On the Doctrines of Hippocrates and Plato 9, but much more important was Galen’s treatise On Demonstration, which was specifically devoted to scientific method, and yet medieval scribes had no interest in preserving it (Nutton 1999: 166, §P.82, 3-5 lists sources containing extant fragments of it, and Hankinson 1991b attempts to reconstruct Galen’s scientific method from his extant works). On Galen’s related interest in mathematics, and mathematical sciences and methods, see discussion in chapter seven of Carrier 2016 and example in section 3.6.VI. For examples of his commitment to an almost modern empiricism see Galen, On the Method of Healing 1.4, 2.7, 3.1, and 4.3 (= Kühn 10.31, 10.127, 10.159, 10.246) and On the Affections and Errors of the Soul 2.3 (= Kühn 5.66-69 and 5.80-90). That Galen’s epistemology was influential in the development of modern scientific method is argued in Crombie 1953: 27-28, 40- 41, 74-84, and Walzer & Frede 1985: xxxiv-xxxvi. I think one could argue the same of Ptolemy’s as well (e.g. consider his anticipations Occham’s Razor in Planetary Hypotheses 2.6 and Almagest 13.2).

[ 809 ] See Kieffer 1964.

[ 810 ] See chapter 2.7, discussion in chapter 1.2.III, and relevant discussion on Galen in chapter seven of Carrier 2016.

[ 811 ] Hero, Pneumatics 1.pr. (see discussion in Argyrakis 2011). Hero also implies here that he had demonstrated other relevant principles in his treatise on waterclocks, which is unfortunately lost. Similar patterns are visible in various works by Galen and Ptolemy (see the end of sections 3.2, 3.3, and 3.4 for examples).

[ 812 ] Ptolemaïs, On the Difference Between the Aristoxenians and the Pythagoreans, frg. 3, quoted in Porphyry, Commentary on Ptolemy’s Harmonics 25.3-26.5. See also supporting quotation of Ptolemaïs in chapter 2.7 and the sources for Ptolemaïs in Carrier 2016 (index).


  1. Can’t wait to get it. The only one I haven’t read yet. Also, have you written anything about debunking the assertion that Judeo-Christian values are the foundation for western society? I hear this all the time and usually say that it’s probably more Greco Roman, but I’m not really sure. Is this addressed at all in the new book? Or is there any chance that you’ll write about this in future?

    Many thanks!


Add a Comment (For Patrons & Select Persons Only)