What do Roman methods of processing, organizing and representing information reveal about Roman institutions and the sociology of knowledge in the Roman world? This complex and ambitious book employs an approach inspired by cognitive science, in particular by theories of distributed cognition, to pose and investigate these questions. After a methodological prolegomenon, Riggsby examines the evidence for Roman use of what he identifies as advanced techniques of information management: ordered and indexed lists (chapter 1), tables (chapter 2), standardized weights and measures (chapter 3), followed by an analysis of spatial depictions in Roman art (chapter 4) and of Roman plans and maps (chapter 5).
In terms of its subject, scope and theoretical framework this is the first study of its kind, and its central questions are manifestly important. For example: one can hardly think of a domain more crucial for assessing the functionality and impact of the Roman system of imperial administration than its ability to process, store and wield information. Riggsby’s creative exploitation of the evidentiary value of different ways of arranging and encapsulating data—e.g. spatial representations as reflections of imperial logic applied to physical landscapes, or listing methods as indicators of the capacity for mobilizing administrative knowledge—showcases the potential of this approach to generate new insights, even from well-known sources. Particularly valuable is Riggsby’s forward-looking discussion of historical problems that may be fruitfully explored from the perspective of information technology (210-222), such as an investigation of Roman archival practices (212) and a broad diachronic examination of the data collected, stored and deployed by Roman administrative entities (215-216). This is a highly promising and exciting new direction for research on the Roman empire.[1]
Riggsby describes the premise of his book as a “culturalist” take that tries to avoid the pitfalls of false equivalence or negative comparison between ancient and modern phenomena and seeks to understand antiquity on its own terms (209-210). This is exemplified especially by chapters 4 and 5, where Riggsby critiques primitivistic assessments of Roman spatial depiction and engages in constructive analysis of what Roman artists and mappers were trying to achieve, how and why. Elsewhere, a wider selection of evidence would have modified the picture: in chapter 2, Riggsby’s narrow definition of tables as matrices of multiple columns designed to be read in both directions (45, 54, 209) leads to a negative discussion of the paucity of such tables in the Roman context. The routine use of two-column tables (e.g. in both public and private accounting) in the Roman empire falls by the wayside,[2] and the existence of complex lists that functionally resemble the advanced tables described by Riggsby is never mentioned.[3] Much of the relevant evidence consists of Greek papyri from the eastern provinces, which brings us to a major problem concerning the evidentiary basis of the book.
Riggsby chooses to limit his study to the “Latin-speaking world” (3-5) and observes a hard distinction between “Greek” and “Roman” practices of managing and representing information. This opposition is puzzling: Greek was used in documents throughout the eastern Mediterranean, from Achaea to Bithynia, Judaea, Mesopotamia and Egypt, and if these regions shared practices in managing information in the Roman period, these commonalities were more plausibly due to their being provinces of the Roman empire than to any shared cultural traditions. Riggsby’s a priori exclusion of Greek-language sources leads him to omit from consideration the most copious and detailed evidence for Roman institutions and administration, documented by hundreds of Greek inscriptions and thousands of Greek papyri that have survived in arid regions of Rome’s eastern provinces (primarily Egypt, Mesopotamia and the Judaean Desert). That the documentary production and record-keeping activity of the Roman state in the eastern empire should be excluded from a study of Roman information technologies is peculiar. Because papyrological material is in many cases the only surviving evidence from the Roman empire that directly speaks to Riggsby’s research questions, its omission has detrimental consequences for the arguments and conclusions of the book. I will illustrate this with a few examples:
One of Riggsby’s main contentions is that advanced methods of organizing written data (e.g. indexed and nested lists) were “fairly rare” in the Roman world and limited to a few narrow contexts in which their function was more about projecting authority than actually locating the data (6-7, 29-41, 80, 203-204). Riggsby also posits a Roman “norm” against obligatory cross-referencing (i.e. having to look up a second document in order to fully grasp the content of the first) which Roman information technologies were supposedly unable to support (20, 32, 204-205, 217).
These assertions are demonstrably inaccurate. As epigraphic and papyrological sources show, the use of indexed and nested lists was a standard practice at all levels of Roman administration. For example: from the earliest period of Roman rule in Egypt, rolls of administrative and fiscal records were numbered and indexed by roll and column (tomos, kolle̅ma), which was the papyrological counterpart of the Roman practice of indexing records on waxed tablets by tabula and pagina/cera.[4] Any information taken from these records was cited by date and roll and column (or tabula and pagina/cera) number. That these indexing marks were not merely rhetorical (“more about authority than recall,” 203) is made clear by evidence for records being routinely looked up in archives[5] and authenticated copies being produced[6] and demanded by Roman provincial authorities.[7] Widespread use of nested lists is illustrated by the vast flow of declarations (census, birth, death, property, etc.) from provincial inhabitants to officials.[8] Of the many different types of declarations involved in Roman administration, only birth declarations are briefly mentioned (15, 30-31) while the Roman census is not discussed at all. And it is mysterious why these forms of documentation are repeatedly described as “rare” when they were a generalized administrative practice throughout the empire.[9]
Riggsby’s claim that Roman information technologies were not sophisticated enough for obligatory cross-referencing is refuted by various sources, including two well-known Latin inscriptions that are analyzed at length in the book. The Trajanic alimenta tablets (discussed at 67 and 70-71) list the names of property holders, with each entry enumerating properties used as real securities for state loans that generated interest for the alimenta program. The minimalistic description of these properties (e.g. “former fundus of Terentianus Paternus with Rutilius Lupus as neighbor valued at 100,000 sesterces“) constitutes an implicit cross-reference to separate registers of landholdings containing more precise information about their location, size, cultivation status, etc. In other words: the alimenta registers provide enough data for the properties to be located in other administrative records but do not replicate the data contained in those records. There is plenty of evidence for this sort of cross-referencing—e.g. census declarations that refer in general terms to properties documented with greater precision in registers of landholdings and urban real estate, whereas contracts of sale typically include this more detailed information[10]—in documents from Roman Egypt.
The omission of key sources leads to other surprising claims, e.g. that the Roman state did not conduct land surveys but was dependent on declarations submitted by individuals, or that centuriation was limited to contexts where land was distributed by the state (169-172, 186-189, 199). Actually, it is well-documented that centuriation was generalized—it is attested throughout Africa Proconsularis, even encompassing the territory of free cities such as Hadrumetum and Lepcis Magna[11]—and that land surveys of different sorts were conducted, as evidenced by African inscriptions concerning the lex Manciana and lex Hadriana de rudibus agris and fragments of cadastral records and dozens of references to routine land inspections (episkepseis) in papyri.
Apart from omission of evidence, there is also a problem of sources being presented in a misleading fashion. To cite only a few examples: we are told that Pliny (NH 25.8) expresses “grave doubts” about illustrations provided by Greek botanical writers (161). In fact, all that Pliny says is that botanical writers had painted the flower of a rare plant yellow, whereas Homer had described it as white (Graeci auctores florem eius luteum pinxere, cum Homerus candidum scripserit). There is nothing to suggest that Pliny doubted the botanical writers had independent knowledge of the plant. And yet, Riggsby proceeds to use this passage as evidence for a general Roman avoidance of illustrations and anxiety about reproducibility (162, 204-206). We are also led to believe that there were “no rules about general usage of measures, or even their original construction” (108) even though the existence of such rules is clearly implied by sanctions against the corruption of official measures mentioned in Roman legal literature. That the rules themselves are not preserved in the Justinianic corpus of Roman civil law is hardly surprising.
Elsewhere, in an effort to show that Roman administration was incapable of achieving any level of standardization of weights and measures (chapter 3), Riggsby states that the Roman modius was measured on five different standards ranging from 8.5 to 13 litres (92). This elides the fact that the different modii are terminologically distinct in ancient sources (e.g. the “flat” or “heaped” modius, the Italic modius and the modius castrensis) and were, as far as the evidence shows, quantitatively consistent. In a similar fashion, Riggsby tries to deconstruct the Roman foot, citing Hyginus on the existence of a “Drusian foot” corresponding to 1 1/18 of a Roman foot among the Tungri in Upper Germany while also noting a difference of 1/24 between the Ptolemaic and Roman foot (92). If anything, this passage underscores the existence of a rather precise notion of the Roman foot, corroborated by a multitude of other evidence. And yet, Riggsby goes on to say that he has demonstrated that “standard’ foot measures do not agree with each other” (115) and that in the Roman context “each user imposes his or her own foot … on the project at hand, rather than being constrained by some external version” (121). Contrary to Riggsby, the financing by local euergetes of public weights and measures does not constitute evidence that these were “local” and “haphazard” (112-113), nor does the coexistence of imperial and local weight standards prove that standardization was “impossible in practice and not even an aspiration of the central government” (127). It is difficult to fathom that in the Roman empire “the bulk of taxes were computed in ways that avoided having to do any kind of standardized measurement” (109) and that officials in Roman Egypt who measured grain to minute fractions of artabai were in each case employing purely local artabai or stylizing approximate figures to give them the semblance of real measurements (122, 126-127).
These examples serve to illustrate how inaccuracies on points of detail throughout the book affect Riggsby’s core arguments and overarching thesis that the Roman informational landscape was irregular, fragmented and effectively “primitive” (8-9, 207-209)—whereas more rigorous and complete examination of the available sources leads to different and often opposite conclusions. When one looks carefully at the evidence, the entire argument of the book becomes less and less persuasive.
Altogether, Mosaics of Knowledge is arguably more successful in its conception than its execution. A combination of capricious selectivity and pervasive imprecision in the interpretation of evidence undermine the broader picture offered in this book. Nevertheless, Riggsby delineates an innovative and productive approach that will surely enrich the study of Roman documentary practices and mentalités.
Notes
[1] It is relevant to mention the reviewer’s own project on Roman archives at the Austrian Academy of Sciences, and A. Dolganov, “Documenting Roman citizenship,” in Ando, C. and Lavan, M. (eds) Roman and Local Citizenship in the Long Second Century, Oxford 2021, 185-228.
[2] See e.g. the tax records in P.Mich. IV and the records of a village administrator in P.Petaus, with masses of other examples. The view of de Ste Croix 1956 (cited by Riggsby at 45 and 63-64) that accounting was typically “in narrative, rather than tabular form” is untenable on the basis of the documentary evidence.
[3] See e.g. P.Fay. 300 descr. = APF 63 (2017) S. 70 Nr. 1 (Arsinoite 1c. CE) enumerating in a non-tabular format several categories of payments under the name of each individual, with totals for each category at the end of the column; there are many examples of this.
[4] See e.g. ILS 5947 (Sardinia, 69 CE); contrary to Riggsby (30 n. 37), Mommsen’s interpretation of the backward C symbol as cera is indeed correct, as confirmed by AE 1999, 571b (Samnium, early 2c. CE).
[5] See e.g. BGU I 73 (Arsinoite, 135 CE); W.Chr. 77 (Arsinoite, 2c. CE); Cod. Iust. 2.42.1 (223 CE).
[6] See e.g. CPR I 18 (Arsinoite, 124 CE), SB III 6995 (124 CE) and SB IV 7362 (Arsinoite, 188 CE).
[7] See e.g. P.Vindob.Bosw. 1 (Arsinoite, 87 CE) and BGU 1 73 = M.Chr. 207 (Arsinoite, 135 CE). Other examples abound.
[8] See T. Kruse, Der königliche Schreiber und die Gauverwaltung, München 2002, I: 63-251.
[9] See the now standard treatment of B. Le Teuff, Census: Les recensements dans l’empire romain d’Auguste à Diocletien, Diss. Bordeaux, 2012.
[10] Compare e.g. BGU I 95 (Arsinoite, 147 CE), a census declaration where an individual registers himself and his family in a “house and courtyard” in the village of Karanis, and P.Lond. III 1164e (Antinoupolis, 212 CE), a contract of sale containing detailed descriptions of two houses, citing their exact location, structure, current and previous neighbors.
[11] See e.g. A. Caillemer and R. Chevallier, “Les centuriations de l’Africa vetus,” Annales 9 (1954) 433–460.