I’ve been reading the French translation by Elisabeth Magnou-Nortier of book 16 of the Codex Theodosianus, the law book compiled under Theodosius II from rescripts or letters issued by preceding emperors. A couple of passages struck my eye, and I translate these here with a few of the footnotes:
Under the empire, custom (mos) continued to play a role as a creator of law. The emperors naturally also created law. The jurisconsult Gaius distinguished three forms of imperial intervention, although there were in fact four in the early Empire.
Edicts (edicta), similar to those issued by the magistrates of the Republic, were formal laws, which had to be posted up in public places. Mandates (mandata) were addressed to functionaries, especially governors of provinces. Decrees (decreta) were judgments given by the emperor in court cases that came before him, either directly or as the last resort of appeal.
Rescripts (rescripta, epistulae) were responses written by the emperor to functionaries who had particular problems on points of law. It has long been thought that “constitutions” were the normal legislative activity of the emperors. But today a re-evaluation has concluded that “the source which is quantitatively the most important of imperial legislation … is not the general law …. but the rescript, a reply from the emperor to a question on a precise point, mostly arising from judicial activity. The editorial methods of the imperial chancellery give these ad-hoc responses a general character; the reply of the emperor never includes … either the name of the person, nor the place concerned. This voluntary reticence contributes to place in high relief the juridical solution, which alone is important.”  …
[Despite the conversion of Constantine to Christianity] material continued to accumulate in the imperial chancellery, often self-contradictory. The rupture introduced by the conversion of Constantine into the legislative activity of the emperors is so perceptible that none of the laws gathered in the Theodosian code is earlier than 312 AD.
The emperors faced a primary difficulty. Since the 2nd century, the jurisconsults had produced a huge volume of legal material and commentary on case law which did not agree among itself. In order to put an end to this endless process, Constantine decided, in 321, to give full authority to the notes of Paulus and Ulpian on Papinian. A still more profound reform was made by Theodosius II and Valentinian III in 426, in a law headed “On Citations”. This law confirmed the authority of five jurisconsults; Papinian, Paulus, Ulpian, Modestinus, and Gaius. If these disagreed, the judge should follow the majority; if they were equally divided, the opinion of Papinian should be used.
An urgent need to intervene was becoming evident. The first collections of laws appear at the end of the 3rd century, with the Codex Gregorianus and at the start of the 4th century, the Codex Hermogenianus which contained the laws of Diocletian. The need for a precise law led Theodosius II to initiate in 429 an ambitious programme, which was cut short. He tried again in 435 in a more modest way. The work was completed on 15 February 438, when the Codex Theodosianus was published in the East. … It was divided into 16 books, the first 5 of which are unfortunately full of gaps where material has not reached us. …
The 16th book has been the subject of discussion by specialists. … Book 16 was composed in the years preceding the publication of the Code.
If we compare the imperial legislation concerning the heresies and sects with that which concerned itself with paganism, a difference of tone is apparent. The law against the first is striking in its density, certainty, indeed a pitiless harshness against those who directly menaced the spiritual unity of the Empire, such as Arianism, Donatism, or Manichaeism. But the laws of chapter 9 show only a limited ambition. They repeat the prohibition of blood sacrifices of animals during ceremonies, first at night and then day and night (364 and 382 AD), because all animal sacrifices for a Christian are as cruel as they are useless. Some prescribe the preservation of temple fabric (399), the upkeep of popular celebrations (399) or again the destruction of rural temples on the condition that this does not provoke trouble or disorder (399).
Certainly a law of 392 indeed prohibited, or more accurately attempted to prohibit, the domestic pagan rites, but we may doubt whether it had any real force. Others order the closure of all the temples, the destruction of sculptured images of the Olympic gods (392 and 395), suppress the privileges and incomes of the pagan priests (396), annex the property of the temples to the state treasury and the income in kind to the army (407). Leaving aside the statue of victory in the Senate … all historians agree that the destruction of the temples, above all in the East, such as Zeus at Apamea in 388, the Serapeum at Alexandria in 391, and numerous temples in Egypt, Phrygia, Cappadocia, and Syria at the end of the 4th and start of the 5th centuries, are the work of enterprising bishops and above all extremist monks. “The destructive zeal of the monks could go well beyond what was authorised, as when they burned in 388 the synagogue of Callinicum.” … The emperors and their counsellors quickly assessed the degree of alienation and resistance by the ancient cults, their rites and festivals, which they judged less threatening to the unity of the empire than the new heresies or the sects, since they were old and above because they could be reused by Christian activities, these offering numerous favourable opportunities to acculturate the new religion in the urban and rural population. However neither magic, nor occultism, nor divination ever disappeared from the socities which called themselves Christian.
 Magnou-Nortier, Le Code Theodosien, Cerf 2002, p. 16-17. This quotation is given from G. Giordanengo, “Le pouvoir legislatif du roi de France”, in Bibl. Ec. des Chartes, vol. 147, 1989, p. 288. Magnou-Nortier gives further references, although not directly to the statement of Gaius.
 Cod. Th. 1, 4, 1 (321).
 Cod. Th. 1, 4, 3 (426).