Galen quotes with approval in both Food and Simples. On the vexed matter
of the primitive wheats and their correct names, Galen (Food 1.13) turns to
On the Materials of Medicine and ļ¬nds that Dioscorides has identiļ¬ed two
kinds of zeia, and compares them with olura, not so much as plants but in
bread and porridge preparations. As plants, they are distinguished by being
either single grained or two-grained. Here, he simply quotes Dioscorides
and does not comment further, possibly because this level of botany (albeit
medical botany) is beyond his competence.
Dioscorides will help us to place Galen and Athenaeus in this difļ¬cult
area of ordering materials that have very technical applications. In his strik-
ing preface to On the Materials of Medicine, Dioscorides sets out certain
principles.31 He aims for a survey that is complete (1) and looks for more than
written sources. He aims for accuracy (both precision and comprehensive-
ness),32 which is lacking, he says, among recent writers (2). Predecessors have
been inconsistent in ordering material, some being guided by properties,
others by an alphabetical order: Dioscorides will group material by proper-
ties (3ā“5).33 Much attention is given to season, parts of the plant and loca-
tion, for which Dioscorides stresses his widespread travels which have given
him extensive opportunities for autopsy. The virtues of Dioscorides are
largely replicated in Galen, where autopsy, reading and seasonal and botan-
ical details are valued. Galen turns to more literature than Dioscorides might
wish: Herodotus and Homer are quoted in the section on primitive wheats,
alongside Theophrastus and Dioscorides (Homer at 1.13.9; Theophras-
tus at 1.13.11; Herodotus at 1.13.12; Dioscorides at 1.13.13ā“17). We might
31 On the preface see especially Scarborough and Nutton (1982).
32 Compare Galenā™s complaints about the doctorsā™ comments on primitive wheats, quoted above.
33 Scarborough and Nutton comment on the manuscript tradition of Dioscorides which sought to
rearrange his material into alphabetical order.
Galen and Athenaeus in the Hellenistic library 83
then place Galen somewhere between the technical author Dioscorides
and the literary author Athenaeus.
Ordering the results of research
Once the results of research have been gathered how are they to be ordered?
Here too, Galen can help us to understand what Athenaeus is trying to do.
In Food, Galen says (2.1) that he has followed a number of earlier works in
dividing his books into cereals and pulses, other plants and ļ¬nally ļ¬sh and
meat. He also follows those who place the most useful cereals ļ¬rst in their
section, and the most nourishing meats ļ¬rst (starting with pork) in their
section.34 Alphabetical order is thus not used, and for ease of reference the
reader is urged to refer to a particular book for a particular class of food.
Within books, related plants are often put together. Thus wild chickling
(see above) follows chickling because (1.27) ā˜the seed is similar to that
of chickling and some people do not think they are separate speciesā™. In
Simples, by contrast, as we have seen, Galen uses alphabetical order in some
books but not in others.
To order by afļ¬nity or alphabetically? Athenaeus and his semi-ļ¬ctional
diners consider both approaches to ordering the vast literary materials they
have been researching. In some books of the Deipnosophistae, the dialogue
form is used and a speaker reports on his reading at some length. Exam-
ples of this, with some listing in addition, may be found in book 14 on
cakes and book ļ¬fteen on garlands. Elsewhere, a list is given: vegetables
and fruits in book 2, breads in 3 (see above), ļ¬sh in 7, meats in 9, cups in
11. Lists often begin with a reference to an authoritative glossary, such as
Callimachus (643e), who, in his Pinax of miscellaneous writings, recorded
books on cakes. At 676f, Ulpian is urged to show more originality in listing
garlands than mere reference to a standard work, the Garlands of Aelius
Asclepiades. These lists in Athenaeus may or may not follow alphabetical
order. They may refer to other orders. They may also consider afļ¬nities.
Because of the grammatical and lexical interests of the work, afļ¬nities are
often determined lexically rather than botanically or zoologically. This is
best seen in the extensive list of ļ¬sh in book 7. On the muros, for example
(312e), Athenaeus quotes Aristotle, in book 5 of Parts of animals,35 who
is reported as saying that the muros ā˜differs from the smyraenaā™. In the
next entry (313a), on the mainides or sprats, Speusippus the philosopher
34 A different order is followed in On the Thinning Diet, where the list begins with the most efļ¬cacious
foods for thinning the humours, namely alliums.
35 A mistake: in fact book 5 of History of animals.
84 j oh n wi lki ns
is quoted for his book Similars, which declares that, ā˜boakes and smari-
dae are similar to the mainisā™. Speusippus is widely used in book 7, and
quotation from his Similars helps to locate the bewildering world of ļ¬sh
with their many local names and apparently countless species. To some
extent the division between zoologist and lexicographer is artiļ¬cial, since
Aristotle, for example, ļ¬nds difļ¬culties with names36 and Speusippus on
the mainis is clearly trying to compare similar ļ¬sh rather than similar
con clusi on
Both Galen as a doctor, then, and Athenaeus as the orchestrator of a literary
symposium have stepped into the very technical areas of botany, zoology
and many others. Galen does not seem to challenge the working methods
of botanical writers such as Theophrastus and Dioscorides in the same way
that he takes on medical writers such as Mnesitheus and lexicographical
writers such as Pamphilus. From a pharmacological perspective, Scarbor-
ough and Nutton judge that Galenā™s catalogues of Simple and Complex
Medicines do not compare well with Dioscorides for accuracy, excellence
and in particular pharmacological properties.37 Again, (if this judgement
is correct) Galen has entered into a technical area in which organisational
skill, botany and medical skills are all needed, and his botany is not strong
enough. Nonetheless, he uses autopsy and travel, as Dioscorides would
wish, along with seasonal and regional factors and is also greatly concerned
with properties or dunameis.
However, what is at issue in this chapter is not Galenā™s technical exper-
tise as a pharmacologist, but rather his approach to ordering the material
he has gathered. One of the striking things about that approach is the
way in which it is marked, as is also the case for Athenaeus, by engagement
with distinctively Imperial intellectual preoccupations, many of which were
themselves vehicles for exploring relations between Greek and Roman cul-
ture, or between Greek past and Greek present. Galenā™s writing includes,
as we have seen, literary sources such as Herodotus and Homer, which puts
him ļ¬rmly in line with the literary preferences of other Imperial Greek
36 See for example Aristotle on crabs, Hist. an. 525a 30ā“525b 6. He considers close resemblance (para-
plesion), names and varieties that are too numerous to describe and too small to have a particular
37 Scarborough and Nutton (1982) 191. They note that Galen did not harmonise the comments from
the many sources he had read.
Galen and Athenaeus in the Hellenistic library 85
authors. It also includes material arising from arguments over Atticism, on
which he has much to say, often of a dismissive nature. Generally he seeks
clarity and easy understanding in place of the complexity and obscurity
that Atticism might foster. At the same time, as we have seen, Galen was
quite capable of checking a term with Aristophanic usage.38 Atticism is also
an issue for Athenaeus, since Ulpian the symposiarch favours pure Attic
precedents (e.g., 94c, 347cā“e, 366a, 368c) while Cynulcus and others argue
for clarity and contemporary relevance, not least in regard to Latin usage.
Ulpian in fact is not simply an Atticist, but a Syratticist, an Atticist from
Syria, on which a number of Ulpianā™s fellow diners make satirical com-
ment. Certainly attention is often drawn to Ulpianā™s poor understanding
of Latin (e.g., 97dā“98e). This anti-Atticism is close to Galenā™s position in
Food, where he notes Latin terms as required. Thus on smyrnion, or Cre-
tan Alexanders, Galen notes ā˜In Rome now, everybody usually calls this
vegetable olisathron, not smyrnionā™ (Food 2.51).
Atticism is one distinctive intellectual concern of the Imperial period.
Fascination with geographical range, which perhaps owes something to
the awareness of the geographical scope of the Mediterranean world as it
is united under Roman rule, is another, discussed above. A third is the
use of Latin texts by Greek authors. Athenaeus cites some Roman authors
(the brothers Quintilii, cited above, seem to be the most recent). Galen
uses no Roman authors in Foods, but does use Sextius Niger and others in
Simples.39 Dioscorides in his proemium clearly had no difļ¬culty in using
Bassus, Sextius Niger and other hellenising Romans. Galen and Athenaeus
are typical of the Greek technical writing of the Roman world both in
engaging with Roman authors but also in holding them at armā™s length, as
a relatively small part of the network of sources they rely on.
Just as important for our cataloguers as these issues of Greek and Roman
cultural interaction was the project of bringing technical writers from a spe-
cial domain to a broader one. Both Galen and Athenaeus read and brought
medical texts, botanists and zoologists into their own texts, though Galenā™s
practice varies according to treatise. These general principles highlight some
of the range of possibilities open to a compiler in the period. Compiling is
not a term of abuse (as it is often applied to Athenaeus), in the mind of Galen
at least, since he, the cataloguer with utility in mind, clearly sees it as vital
for medical practice. Part of his job is to give a comprehensive account as
38 On Galenā™s Atticism see further (among others) Swain (1996) 59ā“62, Wilkins (2003).
39 Nutton (1988) 316 notes that Galen is not quoted in Latin medical sources until the ļ¬fth century ce.
86 j oh n wi lki ns
accurately and as precisely as possible. Galenā™s claim to authority embraces
interpretation of Hippocrates, the best methodology for classifying foods or
drugs, breadth of reading in scientiļ¬c texts, literary texts and lexicographical
texts and autopsy.
The Deipnosophistae seems to have been composed for readers familiar
with these research techniques. Massive reading is expected; consultation of
library catalogues or the pinakes of Callimachus (for example) on cakes or
ļ¬sh is normal practice. There is some, but limited autopsy. There is criticism
of texts in order to establish authenticity (as there is in Galen). When we
consider the presentation of research, Athenaeus seems to subvert expecta-
tions. Alphabetical ordering of massive amounts of data is sometimes used
and sometimes not (as in Galen, who is not himself always consistent even
within one book). Standard works are sometimes welcomed and sometimes
dismissed. The invasion of the catalogue by foreign material occurs on so
many occasions that we should probably explain the phenomenon as part
of the design of the work. Athenaeus takes us by surprise in this respect,
as he does in his setting up Homer as the architect of the symposium and
the founder of the genre of sympotic prose, and Plato as the main threat
to the genre. The Deipnosophistae is a literary product full of paradox and
surprise. This extends to its form as much as its content, which can present
everything from ļ¬sh sacriļ¬ce to the foundation of the Olympic Games by
a cook. That Athenaeus did not seek to set out his researched data as Pliny
does, with sources set out within books ordered by topic, suggests that lexi-
cography and cataloguing were sufļ¬ciently familiar to non-specialists to be
suitable for subversion within the sympotic genre. Such a picture conforms
with Athenaeusā™ apparent subversion of other forms of power and author-
ity, not least that of Larensis the Roman host, who preaches restraint but
presides over a most lavish table in Rome.40
We are told at the beginning of the Deipnosophistae that Larensis has
gathered together diners who are skilled in special areas of paideia. Com-
parison with Galenā™s methods show how knowledge is gathered in this
period. Athenaeus follows much of this method, and gives it a special twist.
This can even be seen in relation to Galenā™s watchword, utility. Athenaeus
engages with utility, while also often subverting it. The Epitome (the sum-
mary version of the Deipnosophistae which is the only text to survive for
books 1, 2 and part of 3) at 1a says that Athenaeus introduces the uses, names
and explanations of all his subject matter. At 185a, Athenaeus declares that
40 See for example Whitmarsh (2000), Wilkins (forthcoming).
Galen and Athenaeus in the Hellenistic library 87
he is going to introduce the most useful parts of the symposium, for which
Homer is the best witness. And at 694c there is a discussion of songs, which
give useful advice for life. These claims to utility are not frivolous. They
help the reader to see that Athenaeus has blended methods that belong to
technical writing into a sympotic form.
Guides to the wor(l)d
Andrew M. Riggsby
Most of the articles in this book focus on the ideologies implicit in particular
means of packaging and processing information; the present essay has a
more technological focus. I do not mean to suggest that the technological
and ideological approaches are, in the last instance, separable, nor even
that such a separation would necessarily be desirable even to the extent
that it might be possible. But, while one might be able to postulate that,
say, British world imperialism and the industrial revolution that gave it
much material support had shared ideological underpinnings, it is clearly
not the case that ideologies can simply produce any resources that might
further them. While I will discuss the role of technology in the broader
culture, I will approach the question ļ¬rst by surveying just what technology
is available. In the ļ¬rst section I will treat several general issues having to
do with tables of contents in general (beginning with their deļ¬nition)
and several commonalities of the four existing Roman examples. The next
sections examine the distinctive uses to which tables of contents are put in
those four texts (taken in chronological order). Then, I return to a more
general issue of how the potential and actual uses of tables of contents
interacted with their functions in individual texts and with each other.
Finally, I will make a few remarks on causal connections between ideology
and technology in this case.
ta bles of cont ents
I begin with two stipulative deļ¬nitions. A ā˜table of contentsā™ (ā˜TOCā™
through most of this chapter) is a summary of the contents of a work
by means of listing its contents in abbreviated form and in the order of
the text. An ā˜indexā™ is a summary list ordered with reference not to the
particular work but to some more broadly available principle (typically
alphabetisation today) and then keyed back to the work by some diacritic
Guides to the wor(l)d 89
feature (typically page or section numbers today).1 A table of contents may
also include such keying, but it is not deļ¬nitionally necessary. I introduce
the distinction primarily to limit the present study to tables of contents;
to the best of my knowledge, and despite more casual use of the term in
discussing several of the texts below, there are no indices in classical Roman
texts. A table of contents or an index can be placed at the beginning or
end of a work (as varies today by national custom) or at the beginning
or end of large segments of an extended work or at both points. In this
last case the point is normally to divide the work up at larger and ļ¬ner
levels of detail (a common approach in modern cookbooks). In principle,
this can be done with many different levels of detail and, given appropri-
ate typographical devices, without creating separate tables for every level
of observation. Such multiple-level organisation will be one of the main
technical issues considered below.
Now, Doody2 has objected to the use of terms such as ā˜table of contentsā™
and ā˜indexā™ as anachronistic in Roman contexts. In particular, it falsely
naturalises centuries of the eventual evolution of the book and of scholarly
methods. The underlying concern here is signiļ¬cant, but need not deter
us from using the modern terminology as long as we keep the potential
problem in mind. Doody, in the speciļ¬c context of Pliny uses summarium
(taken from the author himself ) but neither this nor any other Latin word
I am aware of has classical authority as a general term for the four lists
of contents I will discuss below. Yet, as we will see, they share numerous
similarities, and it is clearly fair to take them as a group. I prefer to use
the modern terms, but to emphasise that the deļ¬nitions offered above say
nothing about use, and more generally to keep in mind throughout the
discussion the issues raised in Doodyā™s paper.
Let me also introduce a distinction between two uses of the table of
contents. (The two are not exclusive, nor, of course, do they exhaust the
possibilities). One is reference. A TOC may be used to allow the reader
to skip more or less directly to a speciļ¬c part of the text to learn about
any chosen topic. The other is segmentation. A TOC may give shape to
a work which the reader may still read linearly. The works under study
here all comprise fairly miscellaneous collections of ā˜factsā™. These need to
be gathered together in bunches to have much importance. The very order
of a text helps create these segments, but they will be clearer to the reader
1 There are other possibilities even beyond replacing numbers with letters or the like. Chapters are
often marked by colour as well as number. Within such a system, one could use variation in intensity
instead of page numbers.
2 Doody (2001) 2.
90 a ndrew m. ri g g sby
if signalled somehow. Thus the TOC helps set up the intended context for
individual observations. A TOC might be judged ā˜usefulā™ by a reader if it
could perform either of these tasks, even if it did not achieve the other.
How common was the TOC in Latin writing? A few manuscripts of
Suetoniusā™ biographies of grammarians and rhetoricians open with a crude
table of contents, a list of the persons treated in the order of treatment
in the text. However, even editors who feel this is ancient assume it is
not Suetonian.3 Nonetheless, the device does clearly exist in classical Latin
texts. It seems, however, to have been quite rare. The individual books of
Columellaā™s On country matters (Rust.) also begin with such lists, although
these can be shown from Columellaā™s own text and the manuscript tradition
to have been duplicated from their grouping at the end of Book 11 to the
beginnings of the respective books by the ninth century.4 The manuscript
traditions of Plinyā™s Natural history (HN) show similar transformations.5 In
addition to the three just cited, the only other examples I can ļ¬nd are in
Scribonius Largusā™ Compounds (Comp.), a mid ļ¬rst-century ce collection
of medical remedies and in Aulus Gelliusā™ Attic nights (NA). One might
imagine, if they were common only a few centuries later (and even readers
at that time were apparently comfortable with inserting their own, as in
the Suetonius example above), that the apparent lack at earlier times is due
to a gap in the evidence rather than actual practice. When Pliny mentions
the device to Titus in his prefatory letter, he seems to allude to its rarity by
citing a now-lost precedent:
And thanks to you no one else will have to read through all this, but, when anyone
wants something, he will seek precisely that and know where to ļ¬nd it. Valerius
Soranus (in Latin, at least) did this before me in the books which he entitled
ā˜Enlightenedā™. (HN pr.33)6
This Soranus appears to be the Q. Valerius Soranus (tribune of the plebs
in 82 ce) cited by, among others, Cicero on linguistic and antiquarian
matters.7 However, while we can point to intervening examples, they are
hardly from prominent texts.
The actual rarity of the table of contents is also suggested by the fact that
not only Pliny, but also the other three surviving authors preface it with an
3 4 Henderson (2002b) 111ā“13.
Kaster (1995) 41ā“2.
5 Pliny: Doody (2001) 3ā“9; similarly on Gellius see Holford-Strevens (2003) 31.
6 Tu per hoc et aliis praestabis ne perlegant, sed, ut quisque desiderabit aliquid, id tantum quaerat et sciat
quo loco inveniat. Hoc ante me fecit in litteris nostris Valerius Soranus in libris quos Ā–poptĀ©dwn inscripsit.
7 Cic. De or. 3.43, Brut. 169; Gell. NA 2.10.3.
Guides to the wor(l)d 91
First, then, I have added [a list] below of what problems the recipes are calibrated
and ļ¬tted to, and have numbered it, so that it is easier to ļ¬nd what one seeks.
Then to the ailments I have added the names and proportions of the medicines of
which the combinations consist. (Scrib. Comp. pr.15)8
But since it often happens that our memory of what we have learned fails, and it
must be frequently revived from notes, I have added an outline of all my volumes,
and when necessary, it will be easy to ļ¬nd what is to be sought in each one and
how each thing must be done. (Col. Rust. 11.3.65)9
Here I have laid out all together the headings which go with each entry so that it
will immediately be clear what can be found in each volume. (Gell. NA pr.25)10
That is to say, the authors do not just supply a list of topics, but also
explain that it is to be used by the reader to jump to a particular topic. Note
also that none of the four even has a word for ā˜table of contentsā™ or even ā˜list.ā™
They all name what is to be listed: capita rerum (ā˜headingsā™ or ā˜chief pointsā™)
for Gellius, argumenta (ā˜outlineā™) for Columella, and indirect questions for
Pliny and Scribonius. One might even go so far as to note some verbal
similarities between the four prefatory passages and their promise of a
Pliny Gellius Scribonius Collumela
quaerat quaeri quaeretur quaerendum
inveniat inveniri inveniatur [reperior?]
subiunxi ā“ subiecimus subieci
singulis libris qui . . . libro ā“ quoque
[indirect questions] capita rerum [indirect questions] argumenta
Now, none of these usages is individually striking given the general sense
of the passages. In the aggregate, however, we might wonder why there is
not more small variation such as Columellaā™s use of reperior for invenire.
It is possible ā“ though certainly no more than that ā“ that all four passages
have a common source. (For a more speciļ¬c argument about the relation
between Pliny and Gellius, see below.)
8 Primum ergo ad quae vitia compositiones exquisitae et aptae sint, subiecimus et numeris notavimus, quo
facilius quod quaeretur inveniatur; deinde medicamentorum, quibus compositiones constant, nomina et
pondera vitiis subiunximus.
9 Quoniam tamen plerumque evenit, ut eorum, quae didicerimus, memoria nos deļ¬ciat eaque saepius ex
commentariis renovanda sint, omnium librorum meorum argumenta subieci, et cum res exegisset, facile
reperiri possit, quid in quoque quaerendum et qualiter quidque faciendum sit.
10 Capita rerum quae cuique commentario insunt, exposuimus hic universa, ut iam statim declaretur quid
quo in libro quaeri inveniri possit.
92 a ndrew m. ri g g sby
Finally, we should note the relative complexity of the syntax of most of
the items in these tables of contents. A great many are indirect questions
(ā˜whether the universe is ļ¬niteā™, Plin. HN 1.2.1), indirect statements (ā˜that a
young vine must be pruned before the winter windā™, Col. Rust. ind. 4.11), or
at least prepositional phrases (ā˜on ancient frugality and on the sumptuary
laws of oldā™, Gell. NA 2.24). Only a few entries are simply nominal. Goody11
has noted that, though in some respects a quite primitive device, the list
is nonetheless a fairly arbitrary one, far from the forms of normal oral
discourse; most distinctively in this respect lists are characterised by their
discontinuity. The Roman tables of contents resist this formalisation, by
pointing to a more normal discursive context. (That book incipits in Latin
take the form of full sentences suggests a similar lack of abstraction.) The
specialised usage has not yet been internalised by our authors.
sc rib on i us l a rg us
Scriboniusā™ Compounds, apparently datable to 44ā“8 ce,12 comprises a sub-
stantial prefatory letter to one C. Iulius Callistus, followed by a table of
contents, then the bulk of the work ā“ a list of formulae for various remedies.
Though it is the earliest and the shortest (117 Teubner pages) of the works
under consideration, it has arguably the most sophisticated table of con-
tents. The entries are brief and normally begin with a prepositional phrase
with ad giving the ailment to be treated; occasionally this is preceded by a
description of the form of treatment (for all headaches, 4; catapotium for
a productive cough, 87). The most novel feature of the TOC is that each
of these entries is then followed by an index number, keyed to the indi-
vidual paragraphs of the main text, each containing a recipe (occasionally
more than one) for the ailment in question.13 Thus this TOC is quite well
adapted to a reference function.
Scriboniusā™ TOC also seems to show two-level organisation. This is
slightly less apparent from the text than the numeration, but would quickly
become apparent to the user. The text itself is organised by symptoms.
Different treatments for the same or similar ailments are generally found
together. Moreover, the text is subdivided by over 100 lemmata at fairly short
intervals.14 The TOC does not show the same subdivision by lemma, but
11 12 Sconocchia (1983) viā“viii.
Goody (1977) 81ā“2, (1986) 54ā“5.
13 One of the manuscripts omits the numeration, but the prefatory epistle (15) explicitly directs the
reader to look for them.
14 Some of the later lemmata for plasters give the type and inventor, but not the function of the
treatment. These may not have added much use value.
Guides to the wor(l)d 93