Summary
It is a point of contention whether the Theodosian Code contains also obsolete constitutions as foreseen for the projected interim code of CTh 1,1,5, or only valid constitutions (with the exception of Book 16). The text of CTh 1,1,6 is unclear in this point and seems to be a mere continuation of the plan of CTh 1,1,5. However, it appears that the first view does not take into account other statements of Theodosius, and that research into particular subjects shows the compilers have rendered a logically consistent review of the law, without superfluous texts. In view of this evidence it is better to assume that also elsewhere in the Books 1 to 15 as a rule only valid laws were included.
1 In 429 the eastern Roman emperor Theodosius unfolded his grand plan for a Code comprising all the general and valid law, providing so a magisterium vitae. He wanted to collect first all constitutions issued since Constantine. Regarding this preparatory collection of constitutions he said in CTh 1,1,5:
Sed cum simplicius iustiusque sit praetermissis eis, quas posteriores infirmant, explicari solas, quas valere conveniet, hunc quidem codicem et priores diligentioribus compositos cognoscamus, quorum scholasticae intentioni tribuitur nosse etiam illa, quae mandata silentio in desuetudinem abierunt, pro sui tantum temporis negotiis valitura.
Although it would be simpler and more in accordance with law to omit those constitutions which were invalidated by later constitutions and to set forth only those which must be valid, let us recognize that this Code and the previous ones1 were composed for more diligent men, to whose scholarly efforts it is granted to know those laws also which have been consigned to silence and have passed into desuetude, since they were destined to be valid for cases of their times only2.
Apparently obsolete constitutions were to be included also. After these were collected and put into a structure, based on the Gregorian and Hermogenian Codes, Theodosius wanted to combine this first code, the two previous ones and the legal writings into a second code which would carry his name. In this second Code there should be no error and ambiguity (qui nullum errorem nullas patietur ambages … omnis iuris diversitate exclusa magisterium vitae suscipiet). Did Theodosius intend the second code to be also only for law experts (scholastica intentio referring to scholasticae, lawyers), that is, that they could sift the outdated laws? It seems that the absence of ambiguity implies that the final code should be cleared of obsolete rules. The 429 constitution also installed a committee and commissioned it with this project. After six years, in 435, a constitution was issued, CTh 1,1,6, giving detailed editorial instructions and appointing new members of the committee. Its text implies that much if not all of the collecting work had been done. The committee delivered a Code in the autumn of 437. It was used as propaganda for the unity of the empire on its presentation at the wedding of Valentinian and Eudoxia. But it did not comply with the final code as envisaged in 429: it contained only constitutions issued since Constantine, i.e. over the period 312-437. It did not comprise the other two codes nor the legal writings. What wanted Theodosius with this actual Theodosian code? Was it the planned first collection, but with more editorial changes, and did he still plan to execute the rest of the 429 project? Or was it a code smaller in contents than planned, but for the rest omnis iuris diversitate exclusa as he had envisaged for the final grand work, thus without obsolete laws, while he had abandoned3 the rest of the 429 project? And if that were the case, would he have openly admitted the failure of the first code? That does not seem probable.
The question comes down to this: Did the Theodosian Code as published include obsolete constitutions as foreseen in 429 for the first compilation, or were these removed by the compilers as was foreseen for the final compilation in that plan, thereby realising, at least for the constitutions issued since 312, the goal of the envisaged second Code? And returning to CTh 1,1,5 and 6: Was there a change of policy in 435 or is CTh 1,1,6 merely the tuning of CTh 1,1,5?
2 The scholarly focus in the 19th century and shortly after was not directed at this aspect. Seeck’s aim was to reconstruct the chronology of the imperial constitutions and beyond that the imperial itineraries. For that reason his interest lay with the constitutions themselves, not with the way the compilers had worked4. Mommsen discussed whether the compilers had to leave out obsolete laws, which, by not being included in the Code, would lose all significance in court, and seems to assume this when he states that the several Easter indulgences are due to chance. But he assumed in any case that the material the compilers had at their disposal was incomplete and disorderly. The question regarding the inclusion of obsolete laws is not definitively answered by him5. Krüger assumed the compilers did not get the allowance to strike obsolete laws, although there were constitutions which, contrary to the imperial order to include all constitutions, were not included in the Code6.
3 The question was not really raised until Archi published in 1976 his seminal work on Theodosius and his Code7. Archi assumed that the committee saw itself placed before an immense task, after having collected all constitutions: to sift the immense mass of legal writings. On the other hand, a speedy result was desirable in order to make better justice possible. The solution which offered itself was to use the results of the design of 429: the collected constitutions. However, the original plan was no longer to be followed. Had the scope in 429 been to transform the collection of constitutions into one without diversitates, for a different goal, namely for practice, and were the texts, otherwise still fully rendered as to the law itself, to be shortened if necessary to the essentials, elucidated and contradictions discarded (‘limitando i testi all’essenziale, e chiarendone i contenuti, anche rinunciando ad eliminare le diversitates’) – by that one could offer the citizens a magisterium vitae –, now the committee of 435 should only trim and abbreviate the texts. No derogated rule was to be eliminated, the users had to find out what was valid. Since the Lex Citandi was included, the users had in any case a method by which to sift the mass of writings in court. Thus the decision of 435 was the only way to salvage what was still possible of the design of 429. So, according to Archi, the compilers executed the first code of the 429-plan, with fewer editorial powers: restricting the texts to their essentials, clarifying their contents, not removing the diversitates. It implied that the obsolete constitutions remained present8.
4 Archi’s book drew attention to the question what precisely the order of 435 held. Opinions are divided and can be arranged in two camps. Thus Matthews assumes, in Archi’s line, that CTh 1,1,6 was merely the continuation of CTh 1,1,5. Until 435 the committee had just collected laws, from 435 they began editing these texts9. The phrase in CTh 1,1,6 circumcisis ex quaque constitutione ad vim sanctionis non pertinentibus solum ius relinquatur (‘after that which does not pertain to the force of the sanction has been taken out of each constitution, the law alone has to remain’) has ad vim sanctionis, which I interpreted as referring to still valid rules10. For Matthews, however, ‘it seems more natural to read this phrase in the law of 435 as a reiteration of the slightly more expansive statement of the same point in 429…’. If one reads it like that, it is not possible to assume the compilers had the authority to remove obsolete laws. The compilers were merely asked to remove the inanem verborum copiam. That excludes that they had authority to eliminate obsolete rules11. On the contrary, they had ‘to include any law that met the stated criteria, obsolete or not’ (cursive by Matthews). Although he later on observes that there are general laws known which were not included, the question whether the compilers collected all laws issued is not discussed by him12. In that view the editing can only concern the first envisaged compilation with the obsolete texts. Riedlberger also rejects the interpretation of vis sanctionis in CTh 1,1,613. In their view the sancienda res of CTh 1,1,5 is considered to be meaning the same as ad vim sanctionis pertinens in CTh 1,1,6, where it includes obsolete laws. It is true that vis sanctionis may indeed be read as res sancienda, ‘the contents of the sanction’, and thus may also relate to obsolete law. As a result of the critique of Matthews and Huck14 I widened the basis for my choice of interpretation by referring to my own research and put it into perspective by remarking that actual research must decide what is obsolete and what not15.
Although it seems natural to assume that the expressions are meaning the same and to assume that Matthews’ and Riedlberger’s view is correct, there is a snag. The basis of their rejection of vis sanctionis as referring to valid law – which it also may mean – is based on this assumption, namely that CTh 1,1,6 is merely an addition to the plan of CTh 1,1,5. That means automatically that obsolete rules were included and that the two expressions mean the same. Their choice between two interpretations is already determined. It is a circular argument.
Then, what if Theodosius had indeed decided to leave out these, as Manfredini suggests: that Theodosius changed his mind and ordered to make a selection16? The order would have either been part of CTh 1,1,6 and left out by the compilers or perhaps been an internal order, finding its basis in CTh 1,1,5. In that case it is equally possible that the two terms point to a difference. Vis means first of all ‘force, power’ and may refer to the the notion that a rule or legal act is binding17, like in … donatum, hoc verbo ea vis continebitur, quam antea scribebamus, …, ut ea, quae ad instructum possessionis vel domus pertinent, tradenda sint in CTh 10,8,118. Here the donation is to have full force, i.e., also the appurtenances to the land or house should be handed over to the donees. Another example of this meaning is in Nov.Theod. 1,5, of 438, where Theodosius rules that no western law posse proferri vel vim legis aliquam obtinere unless he has confirmed it by his own law. Here it concerns a force of the rule which is binding19. A constitution which has been invalidated (infirmata) or lost its validity through non-use (in desuetudinem abiit): has it still vis? Does it therefore not go against the grain to interpret vis sanctionis as applying also to obsolete laws?
Whatever one may think of this, it is clear that in order to get out of this circular argument it is necessary to find proof elsewhere. We need external proof, namely from analyses of parts of the Code. Either these present us with obsolete laws, or they show that no obsolete laws are present, as my research in Food for Rome did where no obsolete laws were found20. Already that research should make one cautious about the interpretation of CTh 1,1,6.
5 It is precisely the question whether CTh 1,1,6 is a mere continuation of the 429 design for the first code and not a change in the plans. After all, we possess but an edited constitution here. CTh 1,1,6 seems at first sight to be an addition to the editing process. It repeats partly CTh 1,1,5 (distribution over titles, chronology). This not only implies that the constitutions had been collected, it also allows for assuming that the collected texts had already been distributed as much as possible over the projected structure of the Code and put in a chronological order, an important part of the editing process, as already observed by Manfredini21. If, however, Theodosius planned in 435 to present the Code at the wedding of his daughter in the autumn of 437, the time for completing the Code would be less than two years. Inserting the Gregorian and Hermogenian Codes might perhaps still be realistic, but inserting the legal writings was not (it took Justinian’s compilers three years for just arranging the writings). If, on the other hand, he still planned to finish the 429 project, the presentation of the Code at the wedding must have meant an intermediate station of this project22. That still leaves open the possibility that Theodosius also changed the design of the second envisaged code. And if the Code we have is indeed not the first Code as projected in the design of 429, which depends for an important part on the question whether it contains obsolete laws in a persistent way, we should try to explore what might have changed or been done differently in 435 to achieve this result. The way to get an answer here is to make an analysis of the legislation on a special legal subject23. Manfredini has already shown the weakness of relying solely on CTh 1,1,5 and 6 for interpreting the nature of the Code. He compares the two orders of 429 and 435 and concludes, that the first by the order to keep the wording of the constitutions primarily served an antiquarian purpose, whereas the second was a repertory meant to serve practice: hence the wider authority to edit the texts. This served the individual constitutions, not the entirety of the imperial law. This seems to exclude the possibility that the compilers could edit the constitutions in such a way as to make them agree with other constitutions with which they otherwise would be in contradiction, or that they could leave constitutions out altogether. However, the reality is different as appears from Nov.Theod. 1 and the protocol of the Senate’s meeting in Rome: the way the emperor speaks leaves no other conclusion than that the original goal was changed into a collection of selected constitutions24. It therefore does not suffice to stick to the mere texts of CTh 1,1,5 and 6, one first must reply to Manfredini’s arguments. Matthews’ lack of specific research and confirmation has indeed been objected against him25.
Recently, also Riedlberger has assumed that obsolete laws were included26. His opinion is, however, more broadly founded than just on the cited texts. He knows that there are many constitutions which have not been incorporated in the Code, for example some of the Sirmondian constitutions; other are attested in inscriptions or papyri. I found in Pharr’s translation about 70 constitutions of which Pharr could not find the text in the Code or Justinian’s Code (Honoré’s lost constitutions; some may of course be constitutions in the lost part of the Code). Riedlberger investigates several explanations for this absence: negligence of the compilers, loss of the excerpt, the original could not be found, the formal criteria were not met, dismissed because it did not answer to the law valid at the moment of compilation, dismissed because it did not meet the generalitas criterium. He himself sees the last explanation as the essential one27. As to negligence, he thinks the compilers were overall careful. Regarding the loss of a text, two situations must be distinguished. One is that part of the Code was lost when Alaric’s compilers did their work. Riedlberger gives the example of Theodosius’ I prohibition of marriage between cousins which may have stood in the fragmentarily transmitted title CTh 3,1228. It is a unlikely example. This title is almost completely transmitted through the Breviary. Are we now to assume that the copy of the Code these compilers had contained all the constitutions but lacked that specific constitution by Theodosius? Precisely between BA 3,12,1 (= CTh 3,12,2) and BA 3,12,2 (= CTh 3,12,3)? Or are we to assume that a transcriber skipped that particular constitution? It is not plausible. – His other example is the sequence of constitutions in CTh 4,6, where he misses a law by Theodosius I. The complete Code did not have this law in the space between CTh 4,6,4 and 529. But this does not seem a case of later loss of text, rather failure to find the text or declining to include it. Riedlberger does not enter into this but mentions the possibility that an excerpt was placed in another title as we would normally expect and may seem therefore not included, yet was included in a missing part (in Books 1 to 5)30. It is possible but not verifiable. The third possibility is that a text could not be found31. Here the availability of the sources comes into play. For recent eastern laws the eastern archives were used, for western constitutions western archives or collections. Since both parts are fairly equally represented, we may assume, so Riedlberger, that the western basic stock (‘westliche Ausgangsmaterial’) was as complete as the east32. This assumption is not nachvollziehbar. It is based on the original proportion of western and eastern legislation about which we do not know anything.
The underlying assumption here is that the Code comprised all issued constitutions, including obsolete ones. Yet that is precisely what is to be substantiated. Likewise it is not clear what Riedlberger wants to prove with his chronological comparison of Constantinian constitutions. Of the 336 ones transmitted, 125 have proposita or lecta, meaning (so Riedlberger) that they most likely come from private collections, evidence that the compilers could not avail themselves of the original sources. Such a proportion makes it not sensible to draw conclusions from the lack of Constantinian constitutions33. His argument is not lucid. Better is his section on constitutions which failed the formal criteria of CTh 1,1,5 and 6: it is only the case with Constantinian constitutions (probably attributable to the presumed used collections). Notwithstanding these criteria we find such constitutions included. Rightly he poses the question, whether the reverse might not have happened too. But there are no errors with post-Constantinian constitutions and it shows the diligence of the compilers34. Apparently this answers his question.
More important is Riedlberger’s discussion of the question, whether constitutions might have been discarded since they conflicted with the operative law. He rejects this. His first argument is that Theodosius ordered in 429 to collect all laws, valid and obsolete. Although he admits that this precept is not repeated in 435, he maintains that the precept to organise them chronologically implies this. Only in this way divergences were visible, divergences which otherwise should not be present35. Riedlberger errs here. It is one thing to discard an obsolete rule or refinement, it is another thing to add to a previous law a refinement or enlargement. For the latter chronology will be an excellent instrument, for the first not. He mentions the laws CTh 16,5,23 and 16,5,27 (on the Eunomians) and these are indeed recalls – but included in Book 16 which is a case apart (see under 8). He also speaks of the countless cases of such obsolete laws: in almost every title of the Code the compilers included obsolete rules36. Is this correct? The Code has 392 titles with contents, take 10% off for the ‘fast jedem Titel’ and there remain 353 titles for Riedlberger to prove his allegation. That is way more than the titles CTh 16,5 on, i.a., the Eunomians, CTh 15,7. Outside of Book 16 he gives as examples of the countless cases two: CTh 4,6, where his analysis of this title is superficial37, moreover, the example is not so randomly chosen as he says38. Further, he takes CTh 15,6,1 and 2, on the maiuma, again with an insufficient analysis. The maiuma was a feast in Egypt with apparently scandalous aspects39. CTh 15,6,1 of 396 (later: CJ 11,46,1) allows the feast as long as it is decent, whereas CTh 15,6,2 of 399 allows the light theatre (again?) with the exemption of the maiuma40. On first sight it looks as if it is first allowed, now forbidden, as Riedlberger assumes41. Yet the summaria on CTh 15,6,1 and 2, a late fifth- or early sixth-century commentary, say otherwise. To CTh 15,6,1 it says that this feast must be celebrated without turpitude (maiuma sine turpitudine celebrandam), to CTh 15,6,2 it says that the maiuma must be allowed (ludorum festiuitatem maiumae relaxandam). Apparently CTh 15,6,2 must be interpreted together with the restriction of CTh 15,6,1: the allowance of entertainment spectacles included the bowdlerised maiuma, not the original. To say that CTh 15,6,1 allows the feast, CTh 15,6,2 forbids it, thus giving us a case of contradicting laws, is wrong and the result of taking the texts only as historical facts, not as elements of a consistent codification42. Appearances may be deceptive. It is necessary to approach the text as a legal text, as accurately as possible, avoiding prejudices (presupposing that there are obsolete texts may preclude the possibility of the opposite), examining it as potential part of a more embracing regulation; and having legal expertise is more than a great advantage. That is also the case where texts do not seem to contradict each other but to repeat another: this does not have to mean that one was superfluous because of that43.
With only two examples, one from Book 16 where we indeed find obsolete laws44, Riedlberger has not substantiated his claim that everywhere in the Code we find obsolete laws45. That contention can only be supported by research in different areas of the Code. But Riedlberger unfortunately applies what he found for Book 16 – restricted, by the way, to the succession sanctions on three sects and apostates – to all the other 15 books.
Theodosius said that the chronology would make clear which laws or rules were outdated by later laws. It means, implicitly, that a contradiction makes this visible. But what about laws, formally still valid but fallen into abeyance, or abolished by custom? How to recognise them? That too is something those who adhere to the inclusion of obsolete laws have to consider.
6 The other group of scholars assumes that CTh 1,1,6, of which we have only the edited version in the Code, implied a policy change: only constitutions and only valid rules were to be included. This is the essence of Gaudemet’s criticism of Archi on this point46. For the first code, no editorial work on the constitutions was foreseen, whereas for the second code far-reaching editorial power was granted in the 429-plan. If the latter code was no longer executed, then CTh 1,1,6 implied a change with regard to the first code of the 429-plan: dropping the second code and an express power to edit the first code. The question then remains: how much editing?
In his article of 1981 Manfredini concluded on the basis of the constitutions CTh 1,1,5 and 6 and Nov.Theod. 1,1 and 3 that the change in 435 of policy included the elimination of obsolete laws. Theodosius boasted to have dispelled the cloud of the volumes by a compendium. That can only mean that he reduced the great quantity of constitutions of his predecessors, which again can only mean a change in policy in 435: from collecting and publishing all constitutions to a selective collection47. He added to this an analysis of CTh 9,34, De famosis libellis. He draws our attention to Valens’ statement in CTh 9,34,8 ‘saepe ut constituimus’, which is hardly covered by the two constitutions of Valens (CTh 9,34,7 and 8, of which 8 should not count). Were they not general? Not found? That is not likely in view of the subject matter and that Valens’ constitutions would have been archived in the eastern chancery. He submits that the compilers made a choice, considering that in this matter four constitutions by Constantine are included and that the same emperor dealt with it also in his famous edictum de accusationibus, known to the compilers (part of it became CTh 9,5,1). Manfredini then proceeds to a careful analysis of the individual constitutions. The first four deal with different aspects of the procedure. CTh 9,34,5 and 6 are by Constans and treat also of libelli in iudiciis oblati. They do not add anything new and may have been included to show that Constans too was combating this crime, unless in CTh 9,34,5 a mitigation is introduced. In the next constitutions the libellus propositus is considered. CTh 9,34,8 is repetitive and perhaps included because of the phrase saepe ut constituimus. CTh 9,34,9 equals the finder of a libellus to its author and the following constitutions work this out: who knows somebody who reads a libellus and does not denounce him, is culpable too (CTh 9,34,10). Manfredini then wonders why the compilers did not include the chapter in the said edict instead of the four Constantinian constitutions, since that chapter comprises their contents succinctly. The reason will have been that the four texts contain more rules issued by Constantine than the edict. Manfredini concludes that on one hand the compilers eliminated outdated and superfluous texts, eliminated errors and diversities, yet on the other hand were restrained in this by their respect for each imperial contribution, even if it did not contribute something new. Justinian’s compilers combined CTh 9,34,7 and 9 in CJ 9,36,2. None of the other constitutions were included48. Manfredini’s cautious conclusion is too modest. CTh 9,34,5 is not repetitive but says, in agreement with the edict de accusationibus, that a libellus famosus must be burned, thus specifying the way of destroying, and this text may have led to CJ 9,36,1, apparently a phrasing taken from Constantine’s edict which Justinian’s compilers may have preferred over Constans’ reference. CTh 9,34,6 goes further than CTh 9,34,1 which safeguards persons accused in libels against calumny, by stating that they are innocent49. CTh 9,34,8 is not repetitive but adds something: where in other texts the libellus is to be destroyed, here any contestatio, i.e. assertion supported by witnesses, evidently to substantiate a libellus, is declared void. Thus Manfredini’s interpretation is fully supported by his analysis of CTh 9,3450.
Another good argument is what Faustus says when offering the Code to the Senate of Rome: ut in unum collectis legum praeceptionibus sequenda per orbem sedecim librorum compendio, quos sacratissimo suo nomine voluit consecrari, constitui iuberet (Gesta Senatus 2) – ‘so that, after the precepts of the laws have been collected in a compendium of sixteen books, which he has wanted to consecrate with his name, he ordered to establish what must be observed in the realm.’ It would be strange if obsolete laws should be observed too, and there is also no mention of the lex posterior rule. Since obsolete laws are not to be followed, the Code contained only valid laws. It is remarkable that this phrase is not discussed by those, assuming obsolete laws were included51.
In my Food for Rome which was about the transportation overseas or overland and processing of foodstuffs for the public distributions in Rome and Constantinople, I had to deal with many constitutions regarding the navicularii (the ship-owners responsible for the transportation of these foodstuffs, mainly grain), the pistores (the millers/bakers who turned the grain into bread), the suarii (the traders in pigs, used for the meat distributions) and several other people engaged in this entire process. There were constitutions which seemed to repeat previous constitutions, or contradict others. But after analysis it appeared that it was possible to distinguish several corporations of navicularii (Africa, Oriens, Tiber), each with their own rules and specialities. E.g., CTh 13,5,8 and 9 seem to cover the same problem, but one is enacted for Spain and the other for Rome, regulating different situations52. As to references to former regulations, there were many but their absence was in general no problem since the later regulation, included in the Code, made clear what was ruled53. Whatever might seem obsolete or superfluous, was not, and the same went for the pistores and suarii54. Hence my assumption that the compilers had intentionally selected, if not already in the beginning, then in any case in 435 only still valid rules and left out those, abolished or superseded by later ones. My observations, which entailed 119 constitutions on the corporations and services in question, next to other related constitutions, spread over at least three Books, provide a broad foundation for my interpretation of CTh 1,1,6 and the Code as such. My forthcoming book The colonate in the Rome Empire, which deals with the coloni censibus adscripti, shows the same picture55.
Honoré has done further research in order to deal with the question56. His resulting judgment is balanced. Some laws, seeming obsolete, on closer sight are not. Others are, but many come from Book 16, where the compilers may scrupulously have included contradictory laws, and in the end the number of inconsistent laws is not so great as expected if the compilers had included all the laws they found57. Honoré also assumed the compilers had the power to make substantive judgments on the laws and so could remove ambiguities, a conclusion which he based on material found58. Matthews dismisses rather quickly this conclusion: one would prefer a factual rebuttal of Honoré’s examples by him59.
Also Liebs assumes the Code was purged from obsolete laws60. Huck has stated the same view in a good exposition on the danger that historians might misinterpret the Code as a complete collection of the Constantinian and later legislation and therefore can be used as such61. Apart from his statistical argument as related below, his refutation bases on the way the texts of the Sirmondian constitutions have or have not been included in the Code. He concludes that overall the compilers took the liberty of expurgating the Code from obsolete laws and consequently warns again historians for assuming that the Code presents a complete picture of the legislation. Book 16 he does not mention62. The criticism of Riedlberger of the views of Huck63 and of Cimma and Banfi on Sirmondiana 3 (who assumed that the compilers left this text out since it was obsolete, but they did not extrapolate this to the entire Code) does not hold64. Finally, there are the some 70 cases where Pharr could not find the law (‘not extant’, see above, under 5). There were perhaps some present in the missing part, but certainly not all. It is another indication that the Code did not, at least in principle, include obsolete laws (except in Book 16).
7 Mommsen assumed that, although the compilers had to include all laws, inclusion was often a matter of coincidence65. Authors like Huck and Riedlberger have also referred to the totality of legislation, to constitutions left out and to constitutions which should have been included, as did Manfredini (the edictum de accusationibus). It is not a moot question. If one assumes that the compilers in principal included all issued general laws, about 3000, as Riedlberger does, the conclusion cannot but be that they did their job well. If one assumes, like Huck and others, that the Code contained only the still valid ones, many must have been left out66. In the first case the underlying assumption is that the emperors issued in the period 312-437 about 3000 general laws, thus on the average 23/24 per year. Is that a realistic figure? (And connected with this: Was it possible for the compilers to collect all constitutions issued in 312-429/437?) Unfortunately the quantitative argument does not give us a straight answer in this respect. Honoré mentions some 29 seemingly obsolete constitutions, apparently overlooked67. This figure is in any case negligible with a presumed total of circa 3000 constitutions (taking the middle between Haenel and Mommsen) in the Theodosian Code passed down to us. Huck mentions the quantitative argument by relating that according to Honoré 60 to 70 constitutions to which reference is made cannot be found. He opines that although a small number may be attributed to negligence of the compilers, such a high figure must mean that the compilers left others out on purpose68.
Apart from the fact that some may have been present in the not transmitted part of the Code, his remarks lead us to the already formulated question: How many general laws in total were issued in the period 312-429/437? There is no figure on that. What we can do is to see if we can find a frequency of general laws for a certain period and extrapolate this for the whole period. Still, this is a far from perfect method. The frequency of laws in the later years points to a very low issuing of general laws (it does of course not touch the issuing of rescripts and all kinds of decisions for special cases, which must have been frequent). For the last decades before the Code (402-437) the average is 12.669, lower than Riedlberger’s 23 or 24. If it were the average for 312-437, it would result in 1,587 original constitutions, about 58% of what the compilers would originally have recovered70. That is of course not realistic. For example, Mommsen has registered for the year 365 130 original (i.e., unfragmented) constitutions. If we would take that as a yearly average, the figure would be 16,380 original constitutions for 312-437, but now we must reduce the period because we know the figure for the years 402-437 where we may expect more or less completeness. That would make 11,700 for 312-401, and the compilers would have recovered about 23%. Both figures will have included outdated laws. The years we took as a basis, 402-437, will not have included outdated laws, or just very few of these, because we may assume that the recent laws were only very incidentally repealed. But the more we go backwards, the higher the chance that laws were outdated and this the more in times of high legislative activity. The high turnout of laws under Valentinian and Valens may have outdated much previous laws. So have the years 363, 362, 361, 360, 359, 358, 357 and 356 resp. 21, 33, 9, 6, 11, 16, 23 and 10 original constitutions; quite a difference compared to 364 with 64 and 365 with 130 constitutions71.
Next to these quantitative extrapolations Matthews and particularly Riedlberger have pointed out that there are laws which, as far as we can guess, are general but not included in the Code. Matthews seems to assume that the Code comprised of all laws, issued since 312. Riedlberger leaves the possibility open that a number of Constantinian constitutions were not retrieved72. Although we do not know whether the compilers knew of these laws, it gives substance to the idea that the Code did not comprise all issued laws.
So, having actually between 3,400 (Haenel) and 2,783 [2,523 (Mommsen) plus 260 transmitted only in Justinian’s Code]73 constitutions in the intact Code, how can one say, the compilers did a good or a bad job in recovering laws? As long as we do not have an idea how many general laws were issued in the period 312-437, any such statement is completely speculative. And Riedlberger’s assumption is fully dependent on the assumption that the Code comprises basically all issued laws, including obsolete ones; an assumption, which is, to say the least, doubtful. And if we restrict ourselves to the two mentioned calculations, one based on a low legislative frequency, the other on a high frequency, the recovery would cover between roughly a quarter and half of all laws issued in 312-437, not more. This is perhaps sobering reading but it is important to realise that, apart from the figures over the last decades before the publication of the Theodosian Code, which are likely to be reliable (the chance that they were invalidated by later laws is low and the archives are less likely to have shown gaps in these years), we do not possess any indications whatsoever about the size of yearly production of general laws74. Gaudemet has found a higher yearly frequency of legislation for 313-323 than for 326-336: 4.5 to 1.3 as regards constitutions for Africa, 13.2 to 9.2 in general; thus, assuming that both periods are comparable, a considerable shift75. Of course it is difficult to base conclusions on these figures: we do not know how much constitutions for those periods were outdated or lost. And there were evidently periods in which the yearly production of general laws was much higher, such as under Constantine, Valentinian I76 and Theodosius. The output varied per year and emperor. What we nevertheless can say is that if the second code would have to include also all obsolete laws, the compilers missed an awful lot. Was that possible? The answer to this depends on the assumptions about the archives they used and the state these were in77. It makes in any case the supposition, the compilers included all or most of the issued constitutions, questionable: the quantity seems too big for that.
8 I already mentioned that in Book 16 we do see contradictions or repetitions which suggest that here a selection according to validity does not seem to have taken place. Honoré thought that Book 16, on religion, was different since this subject contained so much sensitive material, such as the question of Arianism, that the emperor and the compilers may have judged it wise to render merely the existing constitutions and leave it at that to avoid offending the Arian army leaders. Book 16 is indeed different. The question may be posed: was CTh 1,1,5 applying to this book also? Was Book 16 already foreseen in 429? Liebs has suggested that a book on religion was not so unusual since the ius sacrum had always been part of Roman legal literature, notwithstanding that we know of such works not beyond the 1st century AD78. Yet that suggestion is very unlikely79. After the constitutio Antoniniana the new citizens with their own religions were certainly not interested in the old Roman ius sacrum. It is not probable that authors like Ulpian or Paul had included something on the ius sacrum in their commentaries on the Edict (no edict on this) or the ius civile. To suggest that, when more than two centuries later an emperor dedicates a book to religious matters, there is a continuation with the literature on ius sacrum makes little sense. What Theodosius II did was really remarkable and novel. It was only possible after Theodosius I had made Christianity the state religion. There is also in consideration of the text of CTh 1,1,5 no reason to assume that Theodosius planned this in 429 already. It is not likely that the mentioned earlier compilations or the legal writings had a title, least a book on religious law. Thus if the compilers would have followed Theodosius’ precepts, they would not have reserved in their outlay a book on religion. Nor does Theodosius mention this. It suggests Book 16 was a later addition, collected in haste and not examined as to validity, collated to the end. If it could be proved that its texts had already been collected from 429 onwards, it would indicate that such a book was meant to be included from the beginning onwards, or during the work once it was considered necessary or useful. In my opinion it could have been an innovation, probably added in haste after the Council of Ephesus of 431, either just after 431 or perhaps in 43580.
9 To assume merely that the titles conform to the 429 order for the first Code and that CTh 1,1,6 is just the execution of CTh 1,1,5 (which it is in any case not to the full but at the best a partial execution of the first envisaged code), does not bring us further in solving the question about obsolete constitutions, as the above may have shown. What we need is to check in different ways the question of obsolete laws: by analysing other texts like the Gesta Senatus and Nov.Theod. 1, and by doing the hard work: researching whether a title represents a logical whole as Manfredini insists on, or contains duplicates, contradicting and obsolete laws etc. – then, that is what we are really interested in: it answers the questions, whether, implicitly or by another order, the compilers did expunge obsolete laws as projected in the original code, and whether the Code was useful for practice. And if the first possibility is the case, then we may assume that in 435 the compilers went on with the second phase of the 429 plan – viz. removing the obsolete – , but restricted to the imperial constitutions of 312-437. Since in that respect the second code was finished, it could bear the name of Theodosius and with some right the constitution of 429 could be read in Rome’s senate (whereas the constitution of 435 would not be, since it was a mere executory constitution). But also, since the design of 429 was not yet fully executed, the inclusion of this constitution and that of 435 in the Code also becomes explainable.
10 We encounter two approaches in the question of the obsolete constitutions. One is that the Code contained all obsolete constitutions issued in 312-437, or at least should have. Archi assumes this, Matthews bases this on his interpretation of the connection between CTh 1,1,5 and 6 but he has not checked this by any research. Riedlberger bases this interpretation on his research of Book 16, allowing for some loss of Constantinian constitutions. He assumes that what is valid for Book 16, is valid for the entire Code and does not see Book 16 as an anomaly. He sees his two examples outside Book 16 as sufficient to invalidate the assumption that obsolete laws were eliminated81. But as with Popper’s example of the white and black swans, the existence of one black swan invalidates the proposition ‘all swans are white’, but it does not make the white swans black. A better conclusion would have been: apparently in some cases obsolete laws were eliminated, in other cases not. What he brings as additional check for the Books 1 to 15 for this assumption is negligible and superficial, as is his refutation of authors who maintain the contrary. Thus, since both authors do not provide hard evidence for the Books 1 to 15, their arguments remain merely claims (except for Book 16). Moreover, the connection between CTh 1,1,5 and 6 remains unclear. Other statements by Theodosius are not taken into account. Their position needs proof by adducing sufficient examples of obsolete laws. Any new contribution in this direction has also first to answer the rebuttal by Manfredini.
Those who adhere to other approach depart from facts and assume consequently that the Code was purged from obsolete laws, at least in principle, and for some with the exception of Book 16. These authors are Manfredini, Honoré, Huck and Sirks. The references to what Theodosius had in mind in Gesta Senatus 2 and Nov.Theod. 1 indicate that he wanted his Code to be an authoritative collection of valid rules only. CTh 1,1,6 can be read in this sense. These interpretations have not been refuted. Further, their research into legal themes in the Books 1 to 15 (Manfredini for CTh 9,34, Sirks for CTh 13,5 to 9, 14,3 and 482, and Huck for the Sirmondian Constitutions) provide evidence that here we find logical and consistent accounts of a subject matter, without obsolete laws. These researches have not been disproved. As long as that has not been done, the balance sways in favour of the assumption that the Theodosian Code was, apart from accidental inclusions, purged from obsolete laws in the Books 1 to 15.
Does this, by the way, suggest that the Gregorian and Hermogenian Codes contained obsolete rules too? We have no information on this. Since Justinian’s compilers removed all obsolete laws, we have to rely on those few texts transmitted independently, but they do not offer the answer.
Translation The Theodosian Code …, by C. Pharr e.a., Princeton NJ 1952.
It is theoretically also possible that he still planned to execute the rest, but now with an already ‘clean’ collection of imperial constitutions.
O. Seeck, Die Zeitfolge der Gesetze Constantins, ZSS-RA 10 (1889), p. 1-44, p. 177-251; reprinted in: Materali per una palingenesi delle costituzioni tardo-imperiali, [Accademia Rom. Costantiniana], 2, Milano 1983, con introduzione e tavole di raffronto a cura di M. Sargenti; O. Seeck, Regesten der Kaiser und Päpste, Stuttgart 1919 (repr. Frankfurt a/M 1964). In both works the question is not discussed. Rather did Seeck assume the imperial archives were very deficient. As to previous editors Gothofredus apparently assumed obsolete constitutions were included (Codex Theodosiani cum perpetuis commentariis Iacobi Gothofredi, editio nova, Lipsiae 1736, Vol. I, p. ccvx: ‘ergo datum hoc adiutoribus negotium, non tantum, vt leges, qui in abdito erant, ab obscuritatis iniuria vindicarent, et in vnum corpus colligerent, compingerent’, basing himself on Nov.Theod. 1), while Cuiacius did not touch the subject at all. It was not until the first edition in 1824 of the Gesta Senatus and the constitutions later included as CTh 1,1,5 and 6 in the Codex Ambrosianus C 29 inf., that the question could arise. But Clossius, Baudi di Vesme and other editors did not raise the question either. See for this L. Atzeri, Gesta senatus Romani de Theodosio publicando, Berlin 2008, p. 21ff.
Th. Mommsen, Das theodosische Gesetzbuch, ZSS-RA 21 (1900), p. 164. As later in the Prolegomena, Mommsen’s concern was a philologically correct edition of the Code, not whether the compilers had included obsolete laws (although he apparently thought they had not).
P. Krüger, Geschichte der Quellen und Litteratur des Römischen Rechts, Leipzig 1888, p. 285-286.
See the bibliography in A.D. Manfredini, Osservazioni sulla compilazione teodosiana (CTh. 1,1,5.6 e Nov. Theod. 1), in margine a CTh. 9,34 (de famosis libellis), AARC IV (1981), p. 388 no. 4.
G.G. Archi, Teodosio e la sua codificazione, Napoli 1976 (incorporating his Il problema dell fonti del diritto nel sistema romano del IV e V secolo, published in: Giustiniano legislatore, Bologna 1970), p. 35-37; more emphatically on p. 50ff.
J.F. Matthews, Laying down the law, A study of the Theodosian Code, New Haven – London 2000, p. 51, 70-71; p. 60-61 the collection of material.
A.J.B. Sirks, The sources of the Code, in: J. Harries, I. Wood, The Theodosian Code, London 1993, p. 57; also in: A.J.B. Sirks, The Theodosian Code, A study, Friedrichsdorf 2007, p. 151-152.
Matthews, Laying down the law (supra, note 9), p. 65: ‘it seems more natural to read this phrase in the law of 435 as a reiteration of the slightly more expansive statement of the same point in 429’; also p. 290. Similarly P. Riedlberger, Prolegomena zu den antiken Konstitutionen, Nebst einer Analyse der erbrechtlichen und verwandten Sanktionen gegen Heterodoxe, Stuttgart – Bad Cannstatt 2020, p. 147 n. 217.
Matthews, Laying down the law (supra, note 9), p. 62, 65, 291.
Riedlberger, Prolegomena (supra, note 11), p. 195: testifying ‘von einem eigenwilligen und klar falschen Verständnis’.
Matthews, Laying down the law (supra, note 9), p. 65; O. Huck, Les compilateurs au travail: dessein et méthode de la codification théodosienne, in: Société, économie, administration dans le Code Théodosien, ed. S. Crogiez-Pétrequi, P. Jaillette, Villeneuve-d’Ascq 2012, p. 79-127, here p. 86 n. 20. Unfortunately, Huck, p. 79-127, passes over that I explicitly referred on p. 152-153 to my research on the navicularii and gave also the example of CTh 6,23,3 and 4.
Sirks, The Theodosian Code (supra, note 10), p. 155; I did not, nor do I, exclude that in a specific area there might lie an exception. In my view this is the case with Book 16, as already stated in Sirks, Sources (supra, note 10), p. 154.
Manfredini, Osservazioni (supra, note 7), p. 402-404 compares the two orders of 429 and 435 and concludes, that the first by the order to keep the wording of the constitutions primarily served an antiquarian purpose, whereas the second purported a repertorium meant to serve practice: hence the wider authority to edit the texts. This served the individual constitutions, not the entirety of the imperial law. This seems to exclude the possibility that the compilers could edit the constitutions in such a way as to make them agree with other constitutions with which they otherwise would be in contradiction, or that they could leave constitutions out altogether. However, the reality is different as appears from Nov.Theod. 1 and the protocol of the Senate’s meeting in Rome: the way the emperor speaks leaves no other conclusion than that the original idea was changed into a collection of selected constitutions (p. 405).
Vis may of course also refer to actual physical power, or to the force of nature, as in vis tempestatis, but that does not concern us here.
CTh 10,8,1: Imp. Constantinus A. ad Aemilium virum perfectissimum rationalem. Si quando adnotationes nostrae contineant possessionem sive domum quam donaverimus integro statu donatum, hoc verbo ea vis continebitur, quam antea scribebamus,cum adiacentibus et mancipiis et pecoribus et fructibus et omni iure suo, ut ea, quae ad instructum possessionis vel domus pertinent, tradenda sint. Dat. vi Id. Mar. Mediolano Constantino A. III et Licinio III Conss.
Thus also Pharr’s translation, The Theodosian Code (supra, note 2) a.h.l.: ‘the force of the sanction’. However, translating sanctio with sanction is too limited, since the word often indicates in the Later Empire a rule, issued.
Matthews, Laying down the law (supra, note 9) does not take account of this research at all, relying solely on CTh 1,1,6. Huck, Les compilateurs au travail (supra, note 14), p. 86 n. 20 passes over this aspect. Riedlberger, Prolegomena (supra, note 11), p. 195, appears to have seen my reference but did not check it.
Manfredini, Osservazioni (supra, note 7), p. 392.
Elsewhere I have discussed why Theodosius may have let his 429 project peter out: The Theodosian project and the Lex citandi, in: Interpretare il Digesto. Storia e metodi, A. Padoa Schioppa, D. Mantovani edd., Pavia 2015, p. 76-104.
How attractive Matthews’ book is, it does not contain such an analysis.
Manfredini, Osservazioni (supra, note 7), p. 402-405.
Huck, Les compilateurs au travail (supra, note 14), p. 91ff.
Riedlberger, Prolegomena (supra, note 11), p. 191-199.
Riedlberger (supra, note 11), p. 185.
Riedlberger (supra, note 11), p. 186.
Riedlberger (supra, note 11), p. 187.
Riedlberger (supra, note 11), p. 187.
Riedlberger, Prolegomena (supra, note 11), p. 188. According to him this was for Matthews the main cause of absent constitutions, since for Matthews every letter had generalitas character. But he does not cite for this Matthews but L. Wenger, Die Quellen des römischen Rechts, Wien 1953, p. 537. The only instance where Matthews concludes that constitutions were not included is on p. 127-128. There he finds that some of the Sirmondian constitutions were not included and he assumes they were not found, as also in other cases (which Matthews does not mention, p. 65). The arguments involving the Sirmondian constitutions are based on the assumption that this collection was used for the compilation of the Code. However, that is not proven.
Riedlberger, Prolegomena (supra, note 11), p. 189.
Riedlberger (supra, note 11), p. 189-190.
Riedlberger (supra, note 11), p. 190-191.
Riedlberger, Prolegomena (supra, note 11), p. 192. According to him Sirks and Huck are of the opinion that after discarding the obsolete constitutions such divergencies should not be present.
Riedlberger, Prolegomena (supra, note 11), p. 192: ‘Zweitens können wir in fast jedem Titel des Codex Theodosianus beobachten, dass die Redakteure sehr wohl teilweise veraltete, ja sogar später explizit wieder aufgehobene Konstitutionen exzerpierten. Das mehrfache Hin und Her bei den erbrechtlichen Sanktionen gegen Eunomianer ist das augenfälligste Beispiel (→ S. 635); es ist so unübersehbar das Sirks zur Rettung seiner Theorie postulierte, das 16. Buch stelle eine Ausnahme dar, und (nur) dort seien veraltete und/oder aufgehobene Gesetze exzerpiert worden (Sirks, S. 154)’; ‘als ein [cursive by R.] Beispiel für zahllose’. I would of course like to see also the zahllose but am already happy with just 20 or so. Riedlberger exaggerates grossly, to put it mildly.
Riedlberger, Prolegomena (supra, note 11), p. 192. He assumes, p. 98, Valens confirmed for the east CTh 4,6,4 to please Libanios, but it is assumed that Valens just granted an exceptional favour to Libanios. Then he assumes a law by Theodosius I, mentioned in CTh 4,6,6 and which should cancel this suggested confirmation, is missing, but he interprets CTh 4,6,4 insufficiently. He assumes Theodosius II cancels in CTh 4,6,8 a western law, assuming a universal validity of all laws whereas that must be proven. I shall deal with this title in detail in the second edition of The Theodosian Code, A study.
Riedlberger, Prolegomena (supra, note 11), p. 195: ‘angesichts der oben zum beliebig gewählten Titel CTh. 4.6 angestellten Beobachtungen ist der Gegenbeweis bereits geführt’. However, CTh 4,6 was apparently not ‘beliebig’ chosen: CTh 4,6,3 features prominently in his survey of sanctions on succession outside of the legislation against heterodox people, on p. 290-296. One would like to see examples, really at random chosen.
T. Honoré, Law in the crisis of empire, 370-455 A.D., The Theodosian dynasty and its quaestors, Oxford 1998, p. 144; J.D. Harries, «Sacra generalitas», The administrative background to the Theodosian Code, Homenaje to Gonzalez Martinez, ed. R. Bustamante, Madrid 1994, p. 22. In detail on this feast and the legislation about it: J. Caimi Arcadio, Giovanni Crisostomo et la festa di maiuma, Annali della Facoltà di giurisprudenza di Genova, 20 (1984-1985), p. 49-84, who does not take the summarium into account.
Thus Riedlberger, Prolegomena (supra, note 11), p. 194 n. 298 interprets these texts, without taking note of the summaria.
Riedlberger (supra, note 11), p. 194: ‘Kaum schwerer zu übersehen sind die Fälle, in denen zwei Gesetze – womöglich direkt nach einander – einen Sachverhalt diametral entgegengesetzt regeln, ohne dass sich das zweite explizit als Aufhebung des ersten bezeichnen würde (so etwa CTh. 15.6.1 und CTh. 15.6.2298, oder, oben im Beispiel, CTh. 4.6.5 und CTh. 4.6.6)’. And in note 298: ‘Doch angesichts des Hauptsatzes maiumam … denegamus ist es vielmehr die von Sirks vorgeschlagene sprachliche Interpretation, die nicht möglich ist’. However, it is not a linguistic interpretation which should here prevail, but the legal one which I applied. Riedlberger apparently has no sense for the fact that we are dealing here with legal texts, parts of a larger design. What may also hamper his analysis is that he departs from the assumption that texts may be obsolete and in that case he may already have chosen for CTh 15,6,2. It is evident that this precludes any objective analysis of the Code.
Codification of course not taken in the modern sense, but as a consistent collection of valid rules in as far as laid down in constitutions, systematised according to subject or theme.
Apart from the cases Manfredini thought were repetitive, an example is CTh 6,23,3 and 4. CTh 6,23,4 of 437 refers expressly to the constitution CTh 6,23,3 of 432 and confirms the privileges of the decurions and silentiarii, granted by Valentinian III: i.e., those in the east received what had already been granted to those in the west. C. 4 is therefore not a repetition of c. 3, but a separate constitution for the other half of the empire. The lex posterior rule cannot apply here. These constitutions were issued during the compilation process and their being together in the Code represents therefore the official view, at that moment in vigour, as regards the validity of constitutions and as regards their individual validity.
The Code has 392 titles with more or less contents. If we take 10% off to account for the ‘fast jedem Titel’, 353 titles remain where Riedlberger has to prove his allegation. That is way more than the one title CTh 16,5 on, i.a., the Eunomians, or CTh 4,6.
He further deals with laws, left out since they did not comply with the generalitas criterium (p. 199-211), all dealing with theology. This part does not affect the question here.
J. Gaudemet, La formation du droit séculier et du droit de l’église aux IVe et Ve siècles2, Paris 1977, p. 50.
See above, note 16. Manfredini, Osservazioni (supra, note 7), p. 385-428, here p. 389-410, p. 405: ‘Alle luce di queste parole, pronunciate dall’imperatore ex post, le quali commentano il risultato, il prodotto finito, non ci sembra possibile non ammettere che deve essere intervenuta una inversione di rotta rispetto al programma del 534: dall’idea di raccoltare e pubblicare tutte le costituzioni, sembra si sia passati a quella della sintesi, della raccolta selezionata’. In my analysis of Nov.Theod. 1 I did not pay attention to these aspects since I already assumed that the Code contained only valid rules (Sirks, The Theodosian Code (supra, note 10), p. 77-78). But I readily join Manfredini’s argumentation.
Manfredini, Osservazioni (supra, note 7), p. 412-428.
It is actually the anonymous accuser who commits calumny if he accuses an innocent person. Probably the text aims at the accused person who replies that the anonymous accuser is lying (or who replies that the libel is chicanery) and declares that this cannot be calumnious.
Riedlberger, Prolegomena (supra, note 11), p. 149 n. 221 calls Manfredini’s conclusion an ‘erstaunliche Auffassung’ and apparently categorises it among the ‘phantastischen Ideen’; yet he does not disprove Manfredini’s evidence, which deprives his statement of all substance.
Manfredini, Osservazioni (supra, note 7), p. 407-408 mentions the Senate’s exclamation constituum ambiguum removistis. dictum xxiii, but not these words by Faustus. Matthews dedicated a chapter to the Gesta Senatus but passes this point over, expounding on the acclamationes: Matthews, Laying down the law (supra, note 9), p. 30-54; Riedlberger, Prolegomena (supra, note 11) has nothing on this text. Atzeri in her profound analysis of the Senate’s meeting (Gesta senatus Romani, supra, note 4) is not focused on this question, hence she passes it over on p. 142 and 172-174.
The same is visible in the context of geographical and administrative restrictions, where already Mommsen advises to take these into consideration: Mommsen, Das theodosische Gesetzbuch (supra, note 5), p. 164.
For example CTh 13,5,16, which begins with Delatam vobis a divo Constantino et Iuliano principibus aeternis equestris ordinis dignitatem nos firmamus. After this an enumeration of privileges follows. In such a case there is no need to include the Constantinian law too.
B. Sirks, Food for Rome, Amsterdam 1991. Also in Sirks, The Theodosian Code (supra, note 10), p. 152-153.
A.J.B. Sirks, The colonate in the Roman empire (forthcoming).
Honoré, Law (supra, note 39), p. 142-149.
Honoré, Law (supra, note 39), p. 146.
Honoré (supra, note 39), p. 126.
Matthews, Laying down the law (supra, note 9), p. 290: ‘There is no room for the idea that the editors were empowered to make substantive judgments on the the [sic] laws before them, or that they did make such judgments.’, referring to Honoré.
D. Liebs, review of A.J.B. Sirks, The Theodosian Code, A study, Friedrichsdorf 2007, in ZSS-R 117 (2010), p. 516-539, here p. 533: ‘Erst die zweite Kommission habe den Auftrag erhalten, nicht mehr anwendbare Gesetze wegzulassen, was Verf. gut begründet; dem ist jedoch hinzuzufügen, dass sie diesen Auftrag eher vorsichtig, jedenfalls unvollständig ausgeführt hat’. As to the latter, I would not exclude that possibility, there will have been pressure on them in the years 435-437.
Huck, Les compilateurs au travail (supra, note 14), p. 83-98, for which see below under 7.
Huck, Les compilateurs au travail (supra, note 14), p. 91-98.
Riedlberger, Prolegomena (supra, note 11), p. 195 says: ‘Die Behauptung von Huck (2012 [= Huck, Les compilateurs au travail (supra, n. 14)], S. 89), im Codex Theodosianus seien Dubletten und Widersprüche, anders als bei regelrechtem Exzerpieren zu erwarten gewesen, ‘extrêmement rares’, steht in krassem Widerspruch zur Evidenz’. Next he dedicates p. 195-197 to an exposition on the Sirmondian constitutions, which has to serve as rebuttal of Huck. E.g., Huck argues correctly that Sirm. 5 could be placed in CTh 5,9 instead of in CTh 5,7 as Matthews proposes, but that the constitution which allows for the return of foundlings at double the costs of raising them was left out since it contradicted CTh 5,9,1 and 2. Yet this title is a reconstruction from the Breviary and it is therefore possible that the Alarician compilers left it out. Both authors have not seen this. CTh 7 and 8 are examples of the usual Easter indulgencies and it is not remarkable that they were not included. It should rather be researched why the compiler thought these of importance. Sirm. 3 forbids to deal with ecclesiastical matters in a secular court. Matthews thinks it was present in CTh 1,27. Huck correctly opposes the fact that what we have from CTh 1,27 is from the Sirmondian collection and it thus suggests that only these derived from the Code. If placed in the Code, CTh 16,2 would have been a better place. That we do not find it there is because it restricts the granted privileges to religious matters (Huck, Les compilateurs au travail (supra, note 14), p. 93-95). Riedlberger, notwithstanding his harsh critique, deals only with Sirm. 3. He thinks it could have been part of CTh 1,27, considering that in CJ 1,4, where CTh 1,27,2 figures as CJ 1,4,8, eleven texts from the Theodosian Code are found. Even if not all derive from CTh 1,27, CTh 1,27,2 does. This argument is incomprehensible. The remaining ten texts do not relate to the audientia episcopalis (CJ 1,4 deals with various matters). Further, we have no independent transmission of CTh 1,27 except for the Sirmondian collection. It is a case apart, for which the nature of the collection of the Sirmondian constitutions should be considered. As Vessey sets out, the collection was probably added at the end of the 6th century, in Gaul (M. Vessey, The origins of the Collectio Sirmondiana, in: J. Harries, I. Wood (edd.), The Theodosian Code, London 1993, p. 178-199). We do not know when it was made. G. Haenel, Novellae constitutiones, Bonn 1844, p. 423-424, suggested as moment of writing the end of the 6th century or later but did not express a date of collection. Th. Mommsen, Theodosiani libri XVI, Vol. I, Pars prior, Prolegomena, Berlin 1905, p. ccclxxviii, assumed they were collected before 438, and was followed in this. But Mommsen did not give any argument for this and it is equally possible that the collection was made after 438, at some moment in the 6th century, see G. Barone-Adesi, Tradizione ecclesiastica della legislazione tardoimperiale, AARC 10 (1995), p. 391-399, part. p. 396-398. Sirm. 1 is a case apart because the essential part is suspected of falsity. Riedlberger further criticises Huck by stating that CTh 16,9,1 and 2 say the same (Riedlberger, Prolegomena (supra, note 11), p. 197). Yet he did not read the texts carefully enough. CTh 16,9,1 deals with a Jew buying and circumcising a slave of the Christian faith or another sect: the slave is set free. CTh 16,9,2 punishes a Jew for the mere buying of a non-Jewish slave with the confiscation of the slave. If the slave is also circumcised, the Jew is put to death and punished with the price of the slave, while the slave becomes free – conform CTh 16,9,1. There is certainly a difference in contents, which Riedlberger has not seen. The Summaria Antiqua on the other hand see this.
He treats of the arguments of Banfi and Cimma on p. 197. A. Banfi, Habent illi iudices suos, Studi sull’esclusività della giurisdizione ecclesiastica e sulle origini del privilegium fori in diritto romano e bizantino, Milano 2005, p. 174-175, argued that Sirm. 3 of 384 was not included because it contradicted CTh 16,2,12 of 355 and 23 of 376. Sirm. 3 gave next to bishops also other clerics the privilege of the ecclesiastic venue. Riedlberger admonishes Banfi: the more recent would be discarded in favour of the older rules (but what is wrong with that), and on top of this, also Sirm. 15 of 411 has the same rule as Sirm. 3 and was included as CTh 16,2,41. But this is precisely the proof of Banfi’s argument, viz. that the superfluous rule was left out in favour of a more recent one. It is Riedlberger’s objection which does not convince at all. He uses the same argument against M.R. Cimma, L’episcopalis audientia nelle costituzioni imperiali da Costantino a Giustiniano, Torino 1989, p. 104-105, who assumes that Sirm. 3 was not included because it restricted the ecclesiastical venue to ecclesiastical causae and was in clear contrast to CTh 16,2,12 of 455 and further, in more precise terms, with CTh 16,11,1 of 399. Similarly Huck (see previous note). Riedlberger rejects this because this restriction was also present in CTh 16,2,23 of 376 and 16,11,1 of 399. But the latter, a later constitution than Sirm. 3, cited by Cimma, would prevail over the now superfluous and obsolete earlier constitution, and Riedlberger’s objection does not hold. In both cases Riedlberger himself supplies the proof for the theses of the two authors. One wonders, why Riedlberger did not check his argumentation once more before pillorying colleague scholars.
Mommsen, Das theodosische Gesetzbuch (supra, note 5), p. 164.
Huck, Les compilateurs au travail (supra, note 14), p. 83; Riedlberger, Prolegomena (supra, note 11), p. 212, 218.
Honoré. Law (supra, note 39), in his footnotes on p. 144-146.
Huck, Les compilateurs au travail (supra, note 14), p. 83 n. 12. However, several constitutions may have been included in the not transmitted parts.
For the years 402-437, of which we may assume that the laws were both still valid and easily collected completely, we find in Mommsen’s surveys in his Prolegomena (supra, note 11), counting only original constitutions in both Codes, since these are in any case leges generales: 402: 2; 403: 9; 404: 15; 405:17; 406: 17; 407: 12; 402-407: 2; 408: 27; 409: 32; 410: 19; 411: 2; 412: 37; 412: 15; 414: 18; 415: 25; 416: 26; 417: 9; 418: 7; 419: 6; 420: 5; 421: 4; 422: 10; 423: 18; 408-423: 10; 424: 14; 425: 12; 426: 15; 427: 6; 428: 9; 429: 8; 430: 5; 431: 2; 432: 3; 433: 2: 434: 3: 435: 6: 436: 6; 437: 1; 425-429: 19. In total: 455 over a period of 36 years, 12.6 per year.
Actually, over 312-437 a figure of 1600.
Further research is necessary to see whether there are indeed areas with legislation after a certain moment and no legislation before although it must have existed, as can be established for private law subjects in Justinian’s Code. See for the method to ascertain this A.J.B. Sirks, Das Recht der Soldatenkaiser, in: Das Recht der ‘Soldatenkaiser’, Rechtliche Stabilität in Zeiten politischen Umbruchs?, Ulrike Babusiaux, Anne Kolb (Hrg.), Berlin 2015, p. 31-45.
Matthews, Laying down the law (supra, note 9), p. 291; Riedlberger, Prolegomena (supra, note 11), p. 185-211, has been very diligent here, distinguishing errors of the compilers, later loss of texts, texts not found, texts dismissed on account of trivial and formal criteria, texts dismissed because not in accordance with the law as valid, texts dismissed because not general.
Haenel estimated the original number of constitutions, included in the Code, at 3,400, which figure Riedlberger simply uses: G. Haenel, Codex Theodosianus, Bonn 1842, p. XXXIII, n. *: ‘fere 3500 aut 3400 fuisse videntur’; Riedlberger, Prolegomena (supra, note 11), p. 172, bases himself on G. Haenel, Lex Romana Visigothorum, Lipsiae 1849, p. IX, n. *. Riedlberger has evidently not made any calculations himself. On p. 179 n. 280 he proceeds on the basis of Haenel’s figures. As for Mommsen, basing on Mommsen’s list in his Prolegomena under Tempora et Loci, we find 2,523 fragments, deriving from manuscripts of the Code, the Breviary and Justinian’s Code.
The emperors (and their functionaries) issued next to these laws rescripts etc. to individuals on all kinds of matters, which figure may have been high.
On totals of resp. 50-14 and 146-101: J. Gaudemet, Constitutions constantiniennes destinées à l’Afrique, in: Institutions, société et vie politique dans l’empire romain au IVe siècle ap. J.–C., ed. M. Christol e.a., Rome 1992, p. 329-352, here p. 349. We have to remember that Constantine is described as an active legislator, but also that subsequent legislation may have abolished measures by him (unless one assumes that the Code contained also obsolete laws).
See now S. Schmidt-Hofner, Reagieren und Gestalten, Der Regierungsstil des spätrömischen Kaisers am Beispiel der Gesetzgebung Valentinians I., München 2008, on the legislative production of Valentinian I.
On this see particularly O. Seeck, Die Zeitfolge der Gesetze Constantins, ZSS-RA 10 (1899), p. 1-44, p. 177-251 and O. Seeck, Regesten der Kaiser und Päpste, Stuttgart 1919, p. 16.16-23, who assumed that much in the central imperial archives was lost (p. 2.12-18).
(Liebs, Review of A.J.B. Sirks, supra, note 60), p. 523. He cites F. Schulz, Geschichte der römischen Rechtswissenschaft, Weimar 1962, for this, but Schulz does not give references beyond the 1st century AD.
First, the ius sacrum was the specialty of priests and when in the 2nd century AD the Roman religion on the one hand became a formal state religion and on the other hand a private religion amongst many others, the priestly colleges lost importance and disappeared. With them the ius sacrum, whatever there was became part of the law of succession or criminal law, commensurate to their legal importance: such as the texts on the sepulchrum violatum; but also magic books, which were inprobi: D. 10,2,4,1.
See Millar, A Greek Roman Empire, Berkeley – Los Angelos – London 2006, p. 150-151 for Theodosius’ concern in those years (just before the First Council of Ephesus) for unity in theological affairs, and p. 157-160 for his concerns in the period following (431-436). It will have concerned first of all his own pars imperii, but there is no reason to think that the idea of unity for the entire empire lay Theodosius more at heart than unity in his own part. But since the west might drift off theologically, a political dimension was present here too. He may have hoped to cement the west better to the east and into his sphere of influence by legal unity and reinforce the theological cohesion of the west with the east by that as well. See further on imperial ecclesiastical policy, L. De Giovanni, Ortodossia, eresia, funzione dei chierici: Aspetti e problemi della legislazioni religiosa fra Teodosio I e Teodosio II, AARC VI, Perugia 1986, p. 59-76; G.L. Falchi, Legislazione e politica ecclesiastica nell’impero romano dal 380 d.C. al Codice Teodosiano, AARC VI, Perugia 1986, p. 179-212; J. Gaudemet, Politique ecclésiastique et législation religieuse après l’édit de Théodose I de 380, AARC VI, Perugia 1986, p. 1-22.
Riedlberger, Prolegomena (supra, note 11), p. 195: ‘angesichts der oben zum beliebig gewählten Titel CTh. 4.6 angestellten Beobachtungen ist der Gegenbeweis bereits geführt’. If it were so simple.
And also in my forthcoming book The colonate in the Roman Empire.