LEWIS, NAPHTALI, Hadriani Sententiae , Greek, Roman and Byzantine Studies, 32:3 
(1991:Autumn) p.267 

Hadriani Sententiae 

N aphtali Lewis 

H ADRIAN I SENTENTIAE is the title of a curious document 
with a curious history of scholarly acceptance fol­

lowed by disdainful rejection and neglect. The 
Sententiae are preserved as part of a miscellany of texts pre­
sented in parallel columns of Greek and Latin word-for-word 
eq uivalences. The large extan t number of such bilingual word­
lists, known in Greek as Hermeneumata and in Latin as 
Interpretamenta, attests to their popularity as teaching tools in 
Roman times, especially in late antiquity. In addition to those 
reproduced from mediaeval and Renaissance manuscripts in the 
Corpus Glossariorum Latinorum, fragments of fourteen more 
have been found in papyri and parchments of the first century 
B.C. to the seventh century.1 

Ascribed in the manuscripts-falsely, as long established-to 
the grammarian Dositheus, the miscellany in which the Sen­
tentiae are embedded is generally agreed to have the following 
characteristics: (1) "Both the Latin and the Greek texts [were] 
prepared by [an] unknown magister in the early years of the 
third century"2-or, at any rate, in the first half of that century. 
I shall have more to say, presently, about that date. (2) "The 
compiler of the exercises ... seemingly prepared a Greek 
translation for the Latin selections he made, with a Latin version 
for the [other] pieces he included which already existed in 
Greek." (3) "There was early, if not immediate, alteration of the 
Latin version ... to accommodate the Latin to the Greek."3 

1 J. Kramer, Glossaria bilingua in papyris et membranis reperta (Bonn 
1983), with]. Lenaerts' review ChrEg 62 (1987) 241-45, and Kramer, Tyche 5 
(1990) 37ff. 

2 Cf A. Bataillc, RechPap 4 (1967) 168. 

3 A. A. SCHILLER, "Vindication of a Repudiated Text, 'Sententiae et Epistolae 
Hadriani'," in La critica del testa (=Atti, Secondo Congresso Internazionale 
della Societa Italiana di Storia del Diritto [Florence 1971: hereafter 'Schiller']) 
717-27 at 722. 

267 



LEWIS, NAPHTALI, Hadriani Sententiae , Greek, Roman and Byzantine Studies, 32:3 
(1991:Autumn) p.267 

268 HADRIANI SENTENTIAE 

I 

The wide use of these bilingual teaching tools affected the 
transmission of the text in a variety of ways. One that concerns 
us here was long ago noted in RE: "Their vigorous spread 
evoked many alterations or revisions of the original; therefore 
they have been transmitted in several, sometimes substantially 
divergent, versions."4 

This last point is strikingly illustrated by the survival of the 
Hadriani Sententiae in a longer and a shorter version. The last 
serious edition is that of E. Backing, Dosithei Magistri Inter­
pretamentum liber tertius (Bonn 1832), reprinted in Corpus 
Iuris Romani Antejustiniani (Bonn 1841) 193-214. In his 
Praefatio Backing describes the miscellany as consisting "ex 
trivialibus" but containing those Hadrianic items "quae ad 
veteris iuri historiam excolendam gratissimae nobis esse 
debent." But A. A. Pellat, yielding to the growing communis 
opinio of the document's juristic worthlessness, dropped the 
text from his Manuale Juris Synopticum after the third edition 
of 1862. Since then jurists, classicists, and ancient historians have 
almost universally ignored the Sententiae. In Leopold Wenger's 
mammoth Die Quellen des romischen Rechts (Vienna 1953)-
973 quarto pages citing legal and literary texts, inscriptions, and 
papyri-there is no hint of the Sententiae. The rare writer who 
mentions the document has done so only to dismiss it as 
unworthy of notice. 5 The only place where the text is now 
conveniently available is in Corp. Gloss. Lat. III (30-38, 387-90), 
published exactly one hundred years ago. In the longer version 
the left-hand column is in Greek, the right in Latin; in the 
shorter version the positions are reversed. The significance (if 

• J. Tolkiehn, "Lexikographie," RE 12 (1925) 2468. 
5 E.g. Paul Krueger dismissed the Sententiae as "eine Anekdotensarnrnlung 

seichtesten Inhalts" (Geschichte der Quellen und Litteratur des romischen 
Rechts [Leipzig 1888] 252 n.52; [Munich 19122] 285 n.56). Then, at the 
University of Jena in the 1892-93 academic year-that is to say, at the time of 
the publication of his Corp. Gloss. Lat. III-G. Goetz delivered a lecture 
reviewing Divi H adriani sententiarum et epistularum collectionem. After 
outlining the MS. traditon, Goetz ended with an approving reference to the 
legal scholars, notably Krueger, who "narratiunculas ... nunc ... merito 
exclusaverunt" (as 'fontes iuris Romani'). In the lfandbuch der Alter­
tumswissenschaft Schanz-Hosius merely lists the Sententiae under Hadrian, 
without a word of comment or evaluation (TIP [1922] 8, IV2 [1914] 179). 



LEWIS, NAPHTALI, Hadriani Sententiae , Greek, Roman and Byzantine Studies, 32:3 
(1991:Autumn) p.267 

NAPHTALI LEWIS 269 

any) of this difference is not readily apparent. Bataille surmised 
in a 1964 lecture (supra n.2: 165) that in the papyrus fragments 
of these lists the Latin "traditionally" (his word) was in the left 
column because those vocabularies were prepared for teaching 
Latin to people in the Greek-speaking part of the Roman 
Empire. But in nine of the fourteen Hermeneumata found on 
papyrus, most of them published in the years since Bataille's 
lecture, that goal is discernible in the writing of the Latin words 
with Greek characters.6 

In 1971 the late A. Arthur Schiller, in what he intended as the 
first of a series of studies, reasserted (supra n.3) the validity of 
the document, demonstrating that the Sententiae are not mere 
anecdotes, as held by the communis opinio of the preceding 
hundred years, but do correspond in form and content to the 
reports of such hearings before emperors and provincial 
governors that have been brought to light in recent years in 
papyri and inscriptions. In fact, soon after the appearance of R. 
A. Coles' 1966 monograph analyzing those reports of hearings,? 
Schiller turned his attention to the Hadriani Sententiae. 

No doubt owing to its appearance in a remote publication (not 
in L 'Annie philologique till 1978), Schiller's paper escaped the 
notice of most scholars in ancient studies-even jurists. A 
lengthy monograph on the pronouncements of Hadrian pre­
served in Greek 8 does not mention the Sententiae, let alone 
consider them; and only one reviewer took the author to task 
for the omission. 9 As far as I can discover, Schiller's 'vindica­
tion' of the Sententiae has elicited no published comment what­
soever. 

As a next step Schiller proposed (723f) to conduct a computer­
assisted search of contemporary legal and quasi-legal Latin 
texts-e.g. known "constitutions of Hadrian and succeeding 
emperors ... Trajan's letters in his correspondence with Pliny, 
and even Pliny's letters to the emperor." He further announced 
that "the matching [computer] program for the Latin text of the 
Leiden manuscripts of the Sententiae has begun, and the results 
are eminently satisfactory." Schiller's friends and colleagues 
know that, prior to his sudden and untimely death in 1977, he 

6 Kramer (supra n.l) nos. 5-9, Ilff, 15; cf also no. 1. 

7 Reports of Proceedings in Papyri (P.Brux. IV: hereafter 'Coles'). 

8 F. Martin, La documentaci6n griega de la cancellaria del emperador 
Adriano (Pamplona 1982). 

9 ]._ L. Mourgues, Latomus 47 (1988) 896. 



LEWIS, NAPHTALI, Hadriani Sententiae , Greek, Roman and Byzantine Studies, 32:3 
(1991:Autumn) p.267 

270 HADRIANI SENTENTIAE 

invested considerable time and effort in studying this "extraor­
dinary matching of vocabulary and phraseology ... grammatical 
constructions and syntactical usages"; but, even with the 
assistance of his widow, all attempts to find his relevant notes 
and data have, unfortunately, been unavailing. 

II 

Accordingly, this review of the Sententiae must be taken up 
without benefit of Schiller's lost data. Even a quick perusal 
suffices to reveal that the document presents several kinds of 
problems, textual and contextual. Among the former I single 
out the following by way of example; no exhaustive analysis is 
attempted. Arabic numerals refer to page and line of Corp. 
Gloss. Lat. III. 

Title. The short version has simply Hadriani Sententiae, 'Ab­
ptaVOD a1to<pacr£t~ (387.10). The title of the long version is eEOD 
'A8ptaVOD a1to<pacr£t~ Kat E1ttcr'tOA.a~ [sic], divi Hadriani sen­
tentiae et epistolae (31.3ff). Divi, of course, indicates that this 
compliation was made after-probably long after, as we shall see 
presently-Hadrian's death. The word epistolae in the longer 
version of the thirteen Sententiae is explained by the appendage 
of a letter of Hadrian to his mother, inviting her and his sisters 
to join him at dinner in celebration of his birthday.lO This letter 
and two of the Sententiae are missing from the shorter version; 
whether the omission is intentional or accidental is unclear. 

Two-way Translation. The current state of the text has clearly 
undergone many alterations. Most of the time the Greek of the 
Sententiae looks like a translation of the Latin, but there are 
clear or probable instances of the reverse. Here are three such: 
(a) 32.38f. (hOKOU~ 'tOKOU~, inusuras usuras. The Greek is 
sufficiently idiomatic (although a'tOKo~ XPTlcrtC; is more usual), 
the Latin is not. There is no such word as inusuras, which looks 
like a literal (in + usuras), if uninformed, rendering of a + 
'tOKOUC;. The editor of the shorter version simply changed in­
usuras to iniquas (387.51); Backing accepted that change, but in 
a footnote he hesitantly offered iniustas, an obviously easier 
emendation. (b) 33.42. Esau'tTlC;, ex ipsis. Since the required 

10 As M. T. Boatwright points out, AJP 112 (1991) 533 n.74, «Although this 
letter ... mentions 'sisters' of Hadrian, only (Aelia) Domitia Paulina is 
known." 



LEWIS, NAPHTALI, Hadriani Sententiae , Greek, Roman and Byzantine Studies, 32:3 
(1991:Autumn) p.267 

NAPHTALI LEWIS 271 

sense is 'immediately', the Greek is right and the Latin looks 
like a botched translation treating the Greek as two words. The 
edited shorter version has continuo, E~aU'ti1c; (387.55). (c) 33.30 
and 388.24. o:va-, ~£'taKaA£cracreat (middle), revocari (passive). 
The likeliest explanation of the inconcinnity is that the -creat 
ending caused the unwary Latin translator to take the Greek 
infinitive to be passive. 

De Minimis ("non curat lex, at curant philologi"). (a) 34.13ff. si 
qua alia fuissent iusta appears in Greek as d nva aAAa y£,(avtcrav 
8ixata-perfection itself except for the pluperfect verb, where 
the translator-working, obviously, from Latin into Greek­
not only omitted the augment,ll but, in a more important error, 
matched the Latin plural form, overlooking the neuter plural 
subj ect requiring (in Greek) a singular verb. (b) 35.41 f. The 
Greek genitive absolute has produced petente mulieris in the 
Latin. Here the translator was obviously working from Greek 
into Latin. (c) 36.9. While two manuscripts have tva, which the 
sense requires, Corp. Gloss. Lat. prefers those that have-and it 
accordingly prints-pta. This is patent nonsense, but a 
papyrologist docs not have much trouble discerning how iota 
nu could be misread as rho iota, especially if the original iota had 
serifs at its ends or-a common phenomenon-a thickening of 
the ink at the top of the vertical stroke. The plenitude of such 
misreadings is apparent from a simple glance at the tables 
compiled by H. C. Youtie, The Textual Criticism of Documen­
tary Papyri. Prolegomena (=BICS Suppl. 6 [London 1958J 69f; 
second edition, Supp!. 33 [London 1974] 68f). 

Hopeless Confusion? (a) 33.14-17. The Greek has 8ta~OAilv ... 
llv 80UAOt 1tapffi~uvav (in a corrupted form). This appears in 
Latin as crimen ... quod emerui concitaverunt. Indicated by the 
text of the shorter version (see below), the correction of 
emerui to serui (so already Backing) is easy semantically, 
puzzling palaeographically. The short version, instead of 
clarifying, increases the confusion: the petitioner there avers 
(388.1M) that he is securum crimine quod ipsi concitaverunt. In 
the Greek this becomes 8ta~oAilv 80uA(c)ta llv aUWl 1tapw­
~uvav. The word auwt, here dangling without an antecedent, 
presumabl y refers to the slaves of 8ouAla. Its Latin counterpart 

11 In -tcrav iota for epsilon is a common phonetic spelling: cf F. T. Gignac, 
Grammar of the Greek Papyri of the Roman and Byzantine Periods I-II 
(Milan 1976-81) at I 249ff; on omission of augment see II 224, noting 
especially the pluperfect yqovn in P.Mich. VIII 492 (second century). 



LEWIS, NAPHTALI, Hadriani Sententiae , Greek, Roman and Byzantine Studies, 32:3 
(1991:Autumn) p.267 

272 HADRIANI SENTENTIAl' 

ipsi has no such antecedent. Also unconstruable as the text 
stands is the accusative case of ota~oAilv. The best I can do with 
this is to suggest a haplography-(oux) ota~oAilv-which 
would adeq uatel y express the reason for the grievance but 
would hardly render securum crimine. (b) 33.38. The infinitive 
circumventum (esse) is rendered by u1tov08£u8ftvat. To get 
from the Latin for 'defraud' to the Greek for 'adulterate' or 
'corrupt' bespeaks a mental or palaeographical process to which 
I have not yet found-will I ever find ?-a clue. For now my 
bafflement is complete, the more so as the short version 
renders circumventum esse by 1t£ptYEYpacp8m (388.31), impec­
cable both literally and figuratively. 

III 

Examples can be multiplied, but these should suffice as il­
lustrations-a paucis disce omnia. Unless we are to leave it at 
that, however, we must now confront the central question, that 
of the document's authenticity. For that investigation there are, 
as I see it, two essential criteria, form and content. 

Form. As already mentioned, Schiller's paper makes the point 
that the Sententiae conform to the records of hearings before 
emperors and other high officials that have been preserved in 
papyri and inscriptions published in the last hundred years. To 
be sure, only a very few of the extant documents record 
hearings before the emperor; the majority take place before 
provincial officials. But all adhere to certain standard formulas, 
which Coles carefully detailed. 

Thus (a) beginning towards the end of the first century "the 
opening speech in the case, by one of the participants, is 
introduced by a Genitive [Latin Ablative] Absolute construc­
tion with a participle ... the speech itself being given in Oratio 
Obliqua" (Coles 41). The introductory participle is that of a 
verb of saying or petitioning; Coles lists a dozen variants that 
occur in papyri. Of the thirteen Sententiae, six begin with 
petente quodam and four with dicente quodam (or the like). 
Dicente is rendered in the Greek by AEyOV'to<;, the same verb as 
in the hearings in papyri and inscriptions; petente is translated as 
a l:tOllV'w<; , which does not appear in the extant records. Those 
do sometimes have as the introductory participle one or 
another form of aSloco, a common verb in legal texts in 
postclassical Greek. ai1:Eco is a non-technical term conveying the 



LEWIS, NAPHTALI, Hadriani Sententiae , Greek, Roman and Byzantine Studies, 32:3 
(1991:Autumn) p.267 

NAPHTALI LEWIS 273 

same sense, and its use in the Sententiae looks like the work of a 
translator-such as a schoolteacher-more at home in the 
language of literature than that of law. In the shorter version of 
the Sententiae, petente is twice replaced by postulante, which is 
translated once as Utwuv'to<; and once as astOuvw<;. The editor 
of the shorter version obviously regarded the two verbs as 
interchangeable. 

(b) As the record of a hearing continues, "the Genitive 
Absol ute construction ... is never used by the presiding official, 
who now always has an indicative verb.» A,gain from the late 
first century, that verb is most commonly dm:v, "which may 
derive from the use of dixit in Latin commentarii ... The 
construction ... occasionally ... is followed by Oratio Recta.» 
The Sententiae conform exactly: Hadrian's name in the 
nominative is commonly followed by dixit, dm:v. These are 
occasionaly replaced by synonyms: inquisivit, (av)E1~TtTTla£v 
(34.17,49; 389.19), interrogavit, E1tllpw'tllcr£v (35.32,54; 387.17 
and 389.4 have £STt'tua£v, TtPclHlla£v), porrexit, £1tEDwK£V 
(387.48). 

"The d.1t£v -formula is also found together with other indica­
tive verbs. The most obvious of these is cX1t£Kpivu'to, which ... 
however, is never used by the presiding official» (Coles 42ff). 
Here, too, the Sententiae conform, with respondit, (l7t£Kpi81l 
used of the petitioner or his advocate (33.49££, 35.11; 388.37f, 
389.33). 

(c) The decision (KptcrU;) is rendered in Oratio Recta, in the 
Sententiae as in the proces-verbaux analyzed by Coles (49-52). 

Content. Let us turn from language and form to consider the 
substance of each of the Sententiae-to judge, using the 
touchstone of verisimilitude, whether they present us with 
actual cases or with rhetorical fictions, or perhaps with 
clements of both. 

As already mentioned, the hundred-year-Iong rejection of the 
Sententiae stemmed from the conviction that an emperor 
would not have taken personal cognizance of these personal 
grievances and family disputes of ordinary people. This pre­
sumed justification was blown sky-high by the publication in 
1954 of the a1tOKpll-ta'tU (= responsa) of Septimius Severus 
preserved in the now famous Columbia papyrus (P. Col. VI 
123), of which Schiller was co-editor. A few more apokrimata 
have since appeared on other papyri, so that a challenge to their 
authenticity cannot even arise. And they do not deal with 



LEWIS, NAPHTALI, Hadriani Sententiae , Greek, Roman and Byzantine Studies, 32:3 
(1991:Autumn) p.267 

274 HADRIANI SENTENTIAE 

empire-shaking events, or even with mid-level administrative 
matters. They are imperial replies to pleas from ordinary folk of 
no official status or eminence. A certain Artemidorus and his 
associates get only a three-word answer: "Obey the decisons.» 
To a woman whose name is lost in a lacuna the emperor replies: 
"Women are not prevented from borrowing money or paying 
on behalf of others.» The rest are in a similar vein. And we may 
note in passing that the existence of rulings such as these 
exemplify the masses of details that, as Fergus Millar has insisted 
for some time now, the emperor routinely saw to in person. 12 

Now for the Sententiae themselves. Case No. 1 (31.24-44; 
387.11-21). An applicant for military service asks to be enrolled 
in the Praetorian Guard. Hadrian rules: "For the present serve 
in the Urban Cohort, and if you are a good soldier, at your third 
salary payment [i.e., presumably, after a year] you will be able to 
transfer into the Praetorian.» M. P. Speidel informs me (per 
epist. October, 1991) that he knows of no precise parallel for 
such promotion; attested in other sources are promotions into 
the Guard after five years' service in the equites singulares 
Aug usti, or, in the third century, in the legions. 13 On the 
linguistic side, there is an interesting variant: where the longer 
version renders stipendio by O'l'roVlql, the shorter version has 
~ae~ip, another puzzler to be added to those already noted 
above. 

Case No.2 (31.45-32.12, 387.22-33). A petitioner asks to be 
given the congiarium that his freedman would have received 
had he not been condemned to the quarries. The long version 
adds that the praefectus aerarii imposed that sentence upon the 
freedman under the Lex Aelia Sentia, 14 Augustus' famous law of 
A.D. 4 imposing conditions on manumission. Hadrian rules: 
"Why do you seek to ruin [long version: and steal the con-

12 His views are summarized, after earlier articles, in The Emperor in the 
Roman World (London 1977) ch. V. Those views are far from gaining 
universal acceptance. Along with others, I have objected (BASP 13 [1976] 
161H) that Millar exaggerates the degree to which the emperor involved 
himself in the routine nitty-gritty. Also, different emperors no doubt varied in 
the extent to which they delegated and deputized. 

13 M. P. Speidel, Die Equites Singulares Augusti (Bonn 1965) 4. 

14 Secundum legem aeliam sententiam, with the letters ten crossed out, 
possibly by a modern hand. 



LEWIS, NAPHTALI, Hadriani Sententiae , Greek, Roman and Byzantine Studies, 32:3 
(1991:Autumn) p.267 

NAPHT ALI LEWIS 275 

giarium of] a man on whom you are already avenged? You are a 
scoundrel "15_in other words, "Case dismissed." 

Four elements here call for comment. (1) A freedman could 
indeed be condemned to the mines, according to Ulpian, if he 
raised a hand against his patron. 16 (2) The emperor did indeed 
involve himself in rulings regarding condemnation to and re­
lease from hard labor in the mines: Cod. lust. 9.51 contains a 
series of such orders by Caracalla and later emperorsY (3) We 
may not know all the provisions of the Aelian-Sentian Law. If 
this freedman's condemnation was in fact carried out under that 
law, this is new information. (4) A role for the praefectus aerarii 
in that condemnation seems incongruous. Either this is sheer 
fantasy of the compiler of the Sententiae, or there is some 
textual corruption here. In this latter regard, Ulpian's De officio 
proconsulis states expressis verbis (n.16) that provincial gover­
nors (praesides) had this power of condemnation. It should 
therefore be entertained at least as a possibility that praefectus 
aerarii is a corruption of praefectus Aegypti. 

Cases Nos. 3 and 13 (32.13-32, 387.34-46; 36.49-37.15, 
389.46-390.6). These are complaints by a father and a mother, 
respectively, of neglect by a son. The father pleads that he is 
ailing and indigent, having exhausted his own financial 
resources. Hadrian orders the son: "Take care of your father; it 
is for this that he begot you. And see to it that he does not again 
lodge a complaint against you with me." In the other case, the 
mother asks Hadrian "to order my son to give me something, 
since he neglects me." The son, who is present, asserts, "My 
lord emperor, I do not recognize her as my mother.» To which 

15 Improbus in the longer version, impudens in the shorter. 
16 Dig. 37.14.1: patronorum querellas adversus libertos praesides audire et 

non translaticie exsequi debent, cum, si ingratus libertus sit, non impune [erre 
eum oporteat ... quod si manus intulit, in metallum dandus erit. idem et si 
calumniam aliquam eis instruxit. Cf also 37.15.9. 

17 Two such release orders from the prefect of Egypt have been published: 
P.Berl. inv. 8997 (A.D. 139, Latin), a corrected text of Ch.L.A. X 421, published 
in P.Congr. XVIII (1986) II 351-56: "Avidius Heliodorus to .. , [probably the 
military commander). I order you to release Petesuchus son of Petesuchus, 
who was condemned to the alabaster quarry for five years by Petronius 
Mamertinus [v. p.] and has completed his time. Farewell." P.Berl. inv. 11532 
(5 B I 4639: A.D. 209, Greek), the prefect Subatianus Aquila to the nome 
strategos: "r have released Nigeras son of Papirius, who was condemned to 
the alabaster quarry for five years by Claudius Julianus v. p. and has com­
pleted the time of his sentence. Farewell." 



LEWIS, NAPHTALI, Hadriani Sententiae , Greek, Roman and Byzantine Studies, 32:3 
(1991:Autumn) p.267 

276 HADRIANI SENTENTJAE 

Hadrian replies: "If you do not recognize her as your mother, I 
do not recognize you as a Roman citizen." 

The rhetorical and somewhat cryptic quality of this second 
ruling does not dissociate it from the first in its reassertion of 
the duty of children towards their parents. This is in accordance 
both with longstanding Roman law (e.g. Dig. 37.15.1, 9), and 
with the imperial propaganda-on coins as well as in pronounce­
ments and legislation of all sorts-celebrating the virtues 
(benevolentia, clementia, etc.) of the Roman em peror.18 

Case No.4 (32.33-56, 387.47-388.10) is a complaint about 
usurious practice. One of its problems has alrady been men­
tioned (inusuras in 36.38). There are also others-more, in fact, 
than in any of the other Sententiae. Unclear, for example, is the 
complainant's concern in the matter. Was he a victim of the 
illegal practice, or a third party looking to obtain an informer's 
reward? Since his complaint cites plures faeneratores, the latter 
possibility is perhaps the stronger. But the question remains 
open; the text identifies him only as aliquis. 

Unclear, too, are the details of the alleged usury. The long 
version states the rate per hundred denarii; the short version 
has changed that figure to 1,000. In the long version denarios is 
rendered as XPUOlVOU';;, which is surely a wrong translation for 
silver denarii. The short version has ()llvapta. The change to 
Xpuoivou,;; was made, presumably, in the fourth century or 
later, when loans in gold became common.19 

Hadrian's decision reads: "My prefect, vir clarissimus, will 
investigate this matter and will report back to me.» (In the short 
version clarissimus is replaced by eminentissimus, and excutiet 
by iudicabit.) At least two considerable problems lurk here: 

(a) Can we identify the prefect to whom Hadrian refers in 
such matter-of-fact fashion? I can think of two possibilities. 
(There may, of course, be others.) (1) The eminent jurist Julian 
(P. Salvius Julianus) was, in the course of his career, praefectus 
aerarii Saturni and praefectus aerarii militaris under Hadrian. It is 
perhaps no mere coincidence, then, that Case No.2, as we saw 
above, refers to judicial action by the pracfectus aerarii. (2) 
Beginning in the third century-just when this compilation is 
generally agreed to have been made-the praefectus praetorio 

18 Similarly, in P. Enteux. 25-26 fathers appeal to Ptolemy III and IV as 
beneficent rulers to order their maintenance by an abusive son and daughter, 
respectively. 

19 Cf R. S. Bagnall, Currency and Inflation in Fourth Century Egypt 
(= BASP Supp!. 5 [1985]) 49-55. 



LEWIS, NAPHTALI, Hadriani Sententiae , Greek, Roman and Byzantine Studies, 32:3 
(1991:Autumn) p.267 

NAPHT ALI LEWIS 277 

was accorded a judicial competence. This possibility leads into 
the second problem. 

(b.l) In our document the prefect is styled vir clarissimus, the 
exclusive honorific of the senatorial class in the first and second 
centuries. For the top-echelon prefects (Aegypti, annonae, prae­
torio, urbis), who were always members of the equestrian or­
der, the corresponding epithet was vir perfectissimus or emi­
nentissimus (this last substituted for clarissimus in the short 
version, 388.6). The picture for the third century is less clear. 
Taken at face value Cod. lust. 9.51.1 identifies the praetorian 
prefects as clarissimi viri as early as the reign of Caracalla. But 
that designation is suspect for at least two reasons. First, it 
occurs in the caption, not the body, of the constitution; 
therefore its source may be Justinian's compilers rather than 
Caracalla's law. As A. H. M. Jones remarked in a different but 
related connection, "the evidence, mainly laws whose dates are 
dubious, is to my mind too tenuous to justify any but the most 
tentative conclusions."20 Secondly, in a recently published in­
scription (AEpigr [1988] 1051) Ulpian, who was praefectus 
praetorio in 222-223, still bears the lesser designation of 
eminentissimus vir. 

Whether or not the change in epithet began in the third cen­
tury, the definitive change occurred under Constantine the 
Great. In ILS 8938, of 313-317, there are two praetorian 
prefects, the Licinian styled vir eminentissimus, the Constan­
tinian vir clarissimus. Some ten years later-no doubt in a 
reorganization consequent to his defeat of Licinius-Constan­
tine abolished the lesser appellations altogether, leaving only 
clarissimus. Clarissimi of equestrian rank are clearly in evidence 
in the second half of the fourth century.21 From all this 
emerges, once again, the probability that the text of the 
Sententiae as we have it was, if not compiled, at least revised in 
the course of the fourth century. 

(b.2) The Greek text, too, has its problems: clarissimus is 
rendered as E1tl(H1Jl0-ra'to<;. The official term was, of course, 
AaJl1tpo'ta'to<;. as every schoolmaster-indeed, every literate 
person, at least in the eastern half of the empire-surely knew. 
The Greek for perfectissimus was OWO'Tlllo'ta'to<;, never £1tt-. 
We have before us, it seems, the work of an uninformed 

20 The Later Roman Empire (Oxford 1964) III 15 n.51. 
21 See Kramer, Tyche (supra n.1) 42. 



LEWIS, NAPHTALI, Hadriani Sententiae , Greek, Roman and Byzantine Studies, 32:3 
(1991:Autumn) p.267 

278 HADRIANI SENTENTIAE 

person. Yet it is hard to leave it at that. The translator who mis­
translated, or the copyist who miscopied, must have had some­
thing in mind. He presumably understood clarus in a non­
technical sense of 'notable' or 'famous' (el OLD s.v. 6) and 
rendered it accordingly as £1tlcrrlllO<;. The editor of the short 
version of the Sententiae simply cut the textual Gordian knot, 
changing the adjectives to eminentissimus and £~OX6)'t(HO<; 
(again a non-technical translation). 

IV 

As the problems encountered continue to be variants of those 
already examined, it should suffice for our present exploratory 
purpose to summarize the remaining Sententiae cursorily. 

In Case No.5 (32.57-33.25, 388.10-21) Hadrian stresses that 
the attainment of equestrian rank requires not only the stipu­
lated wealth but also a life free of moral or criminal taint. Case 
No.6 (33.26-36, 388.22-29), in which the petitioner asks that his 
father be allowed to return from exile, elicits from Hadrian the 
following response: "Let me look into my records (commen­
tarii), and you be sure to come back to me." No doubt, details 
not stated here-the cause and circumstances of the sentence 
of exile-would be found in the imperial commentarii. 

With Case No.7 (33.37-34.5, 388.30-47) we return to patron­
freedman relations. Here the complainant avers that his freed­
men (plural, number not given), who conduct the business of 
his tabernas (£pyacr'tTlPux), have been pocketing the receipts and 
giving him nothing. Counsel for the defendants replies that, on 
the contrary, everyone of the freedmen has long been con­
ducting the business in the patron's interest and turning over 
the proceeds to him. The short version-longer in this 
instance-adds, "and are still prepared to do so if he would let 
them." Hadrian's ruling suggests his impatience with litigants 
who waste a court's time with such squabbles: "You freedmen 
see to it that he have no complaint. Have common sense." 22 

Case No.8 (34.6-32, 388.48-389.8) shows the greatest textual 
divergence between the long and short versions. Both begin 
with the complainant's statement that his father's property was 
confiscated and he was deprived of the office of decurio. After 
that we have two quite different texts. It is not so much that the 

22 So the short version, which is preferable. In the long version this last 
sentence reads: "But he too has common sense." 



LEWIS, NAPHTALI, Hadriani Sententiae , Greek, Roman and Byzantine Studies, 32:3 
(1991:Autumn) p.267 

NAPHT ALI LEWIS 279 

long version has 36 words and the short only 23, but rather that 
the very words are almost all different, and the edited version, 
even though shorter, reads like an interpretative version of the 
longer. The duality extends even to Hadrian's ruling. In two 
completely different formulations he chides the petitioner for 
neglecting to come forward sooner, the longer version in­
cluding also a reaffirmation of the principle of stare decisis. 

In Case No.9 (34.33-60, 389.9-23) Hadrian prevents a guardi­
an from going through with a manumission detrimental to the 
interests of his ward. Case No. 10 (35.1-40, 389.24-45) involves 
a claim to a congiarium, and revolves around the identity of the 
designated recipient. Again there are substantial textual differ­
ences between the two versions. Hadrian finds against the peti­
tioner. In Case No. 11 (35.41-36.14; not in the short version) 
the mother of an orphan charges that her child's guardian has 
failed to provide sustenance and has even made off with the 
child's congiarium. The guardian alleges a technical impediment: 
the absence of his co-guardian leaves him powerless to act 
alone. Hadrian brushes aside that excuse and orders him to 
perform the duty for which he was appointed and provide 
maintenance proportionate to the available means. It must be 
self-evident that not all such complaints reached the emperor's 
tribunal, but, human greed being apparently eternal, many a 
guardian of orphans must have been tempted toward malversa­
tion of the funds at his disposal. The recently published Babatha 
archive-also Hadrianic in date, as it happens-contains two 
similar complaints against the guardians of an orphan (P. Yadin 
14f). There, too, the mother of the orphan asks that the child be 
maintained in a manner "commensurate with the style of life 
that befits him." 

In Case No. 12 (36.15-48; not in the short version) a father of 
sons who have been selected for military service asks to be 
allowed to look after them lest their inexperience lead them to 
do something napa 'to Ka8ftKOY, extra ordinem. He is pre­
pared to go even as their servant. Hadrian, apparently touched 
by such devotion, rules: "May the gods forbid that I make you 
subservient to your sons ... You shall be their centurion." This 
episode smacks more of drama than of history; like Case No. 
13, already reviewed, it provides an instance of Hadrian's hu­
man qualities as a benevolent ruler. Yet it could have happened: 
the e.mperor surely did not lack the authority to make such 
appointments. 



LEWIS, NAPHTALI, Hadriani Sententiae , Greek, Roman and Byzantine Studies, 32:3 
(1991:Autumn) p.267 

280 HADRIANI SENTENTIAE 

This rapid survey points, I think, to the following assessment 
of Hadriani Sententiae. If authentic, they have suffered severe 
textual corruption. If rhetorical invention, they neverthless in­
corporate authentic elements, both formal and substantive. 
Enough has been said here, I think, to illustrate the kinds of 
problems that this puzzling document poses. Enough has been 
said, I hope, to convince the reader that those problems de­
serve to be addressed rather than ignored. 

CITY UNIVERSITY OF NEW YORK 

March,1992