The now-deserted site of the city commonly known as Old Sarum (although not to universal satisfaction, for it only acquired 'Old' after New Salisbury had come into being) began life as a massive Iron Age hill-fort, itself raised above the lowlands of Salisbury Plain atop a spur of high ground known as Castle Hill, which lies across the confluence of the Bourne and Avon rivers. There was some presence there of the conquering Romans, for the hill-fort became a nodal point in the system of Roman roads, notably one running from London to Exeter. Though it is still poorly understood, an at least modest and prosperous settlement grew up, mostly situated below the hill-fort and particularly around the London-Exeter road. This persisted into the sub-Roman period; as a stronghold and refuge the hill-fort itself became more intensively settled in the Romano-British period. The Romans had named it Sorviodunum, which is derived from Celtic terms. The Saxons corrupted the first element of the name and converted the second into the equivalent Germanic term, burh, so that it became Searobyrh, while in Domesday this was further corrupted to Sarisburia (our Salisbury) and abbreviated to Sarum.

The name Salisbury was also applied to a much larger estate belonging to the bishopric of Ramsbury and Sherborne and then its successor, the bishopric of Salisbury, which was holder at the time of Domesday. This estate, initially perhaps a single manor but subsequently divided into multiples, encompassed Old Sarum, the future site of New Sarum (modern Salisbury) about two miles south, and all the land in-between; this is the source of some confusion, as it leads to the argument that it was the Salisbury estate that was actually 'old' (and, indeed, so described), relative to the borough atop the hill-fort. Domesday does not identify Old Sarum as a borough but suggests it might have been so – or at least an area reserved, because of its military value, from the royal grant of the episcopal estate – because each third penny of its revenues was due the king.

The Saxons did not occupy Old Sarum after ousting the British, preferring to establish their own centre of power at Wilton, but in 900 it was part of the estate with which the bishopric of Sherborne was endowed. Faced by the Viking incursions of the 870s, however, and after some loss of confidence in the defensibility of Wilton, Alfred had issued orders that Old Sarum be urgently refortified; though the authenticity of this document has been questioned, it is consistent with the general defensive thrust at that period. But there is no reason to think it was urbanized at that time, even though a church dedicated to St. Mary seems to have been built there at some point during the Saxon period, and the Burghal Hidage shows that attention was refocused on Wilton during the main burh-creation period. There was a similar reversal in 1004, when the mint at Wilton, sacked by the Danes the previous year, was transferred to Old Sarum, which had managed to avoid the fate of Wilton; it remained there throughout the century, even after Wilton was rebuilt and moneyers began operating there once more, and moneyers continued to work at Old Sarum down to the reign of Henry II. The transfer was, again, an emergency response and need not imply urbanization at Old Sarum nor, initially, the presence of a market; but associated transfer of administrative functions from Wilton could have attracted, first, the establishment there of a service population and, second, the attention of the Normans, who liked either to draw focus away from or to over-awe Saxon centres of administrative power, where possible, just as they liked to take advantage of existing fortifications during the phase of consolidating their military dominance over newly-conquered territory. The continuation of the Old Sarum mint longer than any other of the county, except that at Wilton, is, on the other hand, suggestive of the presence, following the Conquest, of some administrative and/or commercial activity, both of which had need for coin.

It is therefore no surprise to find the Normans, within a few years of the Conquest, building a timber castle atop the hill-fort, roughly in the centre of its enclosure; this motte and bailey castle, accompanied by an outer bailey and royal residence, and with keep rebuilt in stone ca.1100, was in the 1130s strengthened and expanded for Henry I by Bishop Roger le Poer, with a curtain wall ringing the mound-top, and a luxurious new palace, though Henry did not live to see it finished. Not long after the erection of the original castle, a Church council decreed (1075) that episcopal seats should be in cities; so that based at Sherborne/Ramsbury was transferred to Old Sarum, and Bishop Osmund, kinsman of the Conqueror, completed the construction of a modest cathedral, dedicated to St. Mary, in the north-western quadrant of the site, next to the castle; it was consecrated in 1092. The editorial comment of chronicler William of Malmesbury on the transfer suggests that the site was not perceived at this time as a civil settlement. The cathedral was rebuilt by Bishop Roger (prior to him starting work on the castle), on the kind of impressive scale he favoured, and characteristically he added an episcopal palace on the side of the cathedral away from the castle.

The combination of castle, cathedral, and residences of secular and ecclesiastical authorities (the palace serving the sheriff when the king was not in residence) must have been a powerful magnet for settlement of those who felt they could make a living supplying necessary goods and services, and Bishop Roger – so closely involved with other urban development in the county – would well have understood the desirability of fostering such settlement within the large space of the hill-fort. It seems likely that space allocated for houses would not have taken long to find tenants. Though hard to judge when this process of urbanization began, the third penny in Domesday and the translation of the see suggest that, if not already underway, there was at least an intent within the Norman leadership to urbanize Old Sarum. Roger le Poer would certainly have been capable of following through on such a plan to convert the burh into a borough, or of giving shape to a process that had already begun naturally. The grant, attributed to Henry I in John's confirmation, of a merchant gild and exemption from tolls and other marketing costs (e.g. stallage) throughout the realm, is further evidence of this plan – indeed, a recognition of its fruition. The grantor assumed, based on an understanding of the situation at Winchester, that the Old Sarum burgesses – the trading component of the civil community, who would most benefit from toll exemption – were already, or would become, members of such a gild; it is a reasonable assumption, though we do not know the truth of the matter.

The pipe roll of 1130 mentions tolls from the market at Old Salisbury as having been given by Henry I's queen to its cathedral; that it had previously been accounted for via the farm of Wilton is a further indication of the past close connection between the Saxon county town and the hill-fort. When the pope confirmed this gift in 1146 he mentioned, among the many possessions of the cathedral, not only the market tolls and profits from a market court, but also a week-long fair at Salisbury around the festival of the Nativity of St. Mary (September), and lands which Alward and Godo once held in the borough of Salisbury. It has been long debated whether the civil population of this borough actually lived atop the hill-fort, in the areas not occupied by the monumental structures, or whether only a small number of townsmen, closely involved with the operation of castle and cathedral, had habitations there, with the others below the slopes in suburbs to south, east, and west – the last being remembered in the place-name of Newtown Westgate – known from both medieval and later documents and modern archaeological evidence.

The descriptor of 'suburb' has been criticized on the belief that these extra-mural settlements were really the borough proper, with any burgess presence on the mound-top limited to service personnel associated with castle and cathedral; this was based on archaeological examination of only a small sample of the non-monumental space. In the latter half of the twentieth century the dominant interpretation was that most townspeople settled outside the east gate of the enclosure; here, where ancient roads converged, could have been a plausible site for a market to be held, and archaeology points to a sizable settlement flourishing throughout the Middle Ages. The western suburb would, another theory argued, have been a more convenient base for those servicing the ecclesiastical community and visitors to the cathedral; the size, character, and even the date of this 'suburb' has yet to be investigated by any large-scale archaeological work.

The older opinion that the mound-top was a bustling city was argued again by Daphne Stroud ["The Site of the Borough at Old Sarum 1066-1226: An Examination of Some Documentary Evidence," Wiltshire Archaeological and Natural History Magazine, vol.80 (1986), pp. 120-26], and in 2014 an archaeological survey, using a range of non-intrusive scanning technologies, produced confirmation, in the form of detected subterranean remains, of a fairly extensive network of houses, streets, and industrial features throughout the quadrant of the outer bailey south of the castle, close to the principal medieval gateway at the south-east side of the plateau. Whether this urban area was the result of planned or organic settlement, and whether it incorporated a marketplace, have yet to be determined from the new data; space issues aside, the bishop may have preferred to have the marketplace elsewhere than in the outer bailey, where one-third of its profits would have been due the king (until the queen surrendered them to the cathedral). We should certainly think in terms of urban settlement on the mound-top overflowing to areas below the slopes, perhaps long lightly settled. This may have been happening as early as the time of Bishop Roger, a likely candidate for founding an extra-mural new town in one or more areas where there was already spill-over, perhaps associated with churches heard of but whose location is uncertain, one of them having a dedication (St. Etheldreda) pointing to a Saxon foundation. The area west of the hill-fort incorporates at least one group of plots that appear the burgage type. Yet there is no need to limit our concept of the borough to just one of the east or west end settlements, which would have met needs that were not only residential but also for cemeteries (partly confirmed by archaeology) and for any industrial processes requiring ample water supply or being too noxious to be acceptable within close proximity of castle or cathedral.

With its steep slopes and commanding views, Old Sarum was a good refuge and defensive fortification. But as a long-term centre of population it proved inhospitable: windswept in winter, shadeless in summer, and without a readily accessible source of water (so that high costs were involved in importing it). Such were among the problems that doomed the town atop the hill-fort to a gradual decline. Friction between the clergy and the castle garrison – the castellans controlling the gateways to the site and therefore being able to interfere with freedom of movement of the clergy and with access by worshippers from around the region as well as by pilgrims – was another cause, while the climb up to the hill-fort was also rolled out as further grounds for a divorce. Relocating the cathedral seems to have been mooted by Bishop Herbert Poore, but possibly the idea, and certainly its realization, belong to his brother, the dean and his successor as bishop, Richard Poore. A suitable site within the estate of Salisbury, closer to the river and where there were existing settlements, was within sight from the top of Old Sarum. The arguments for moving were laid out to win approval from pope and king, and, these obtained, in 1219 a churchyard was consecrated and a wooden chapel erected at the future New Salisbury. Although initially left standing in Old Sarum, the bishop's palace was assigned to be demolished in 1237 and the cathedral gradually taken apart over the course of the next century – its foundations rediscovered by archaeologists in the early twentieth century.

The 1220s saw the evacuation of the ecclesiastical community, the most important client group for local traders and artisans, and, finally, the symbolic move of the episcopal tombs to their new home (1226), where new markets and fairs drew away business. Many, though not all, of the townspeople also moved out. It may be that for some time their market had not been doing well, judging from taxation assessments in Henry II's reign, and the mint had been lost earlier. The burgesses had found the money to purchase confirmations of their existing borough liberties when John took the throne, and again upon Henry III's accession, but lacked the means or the inclination to negotiate additional privileges. The grant of a fair in 1246 and reductions in taxation assessments in the 1260s may have been efforts to try to halt the slide; whether the September fair was still being held is unknown. In 1275 the men of Wilton, engaged in a trade war with New Salisbury, complained that the licensed Tuesday market at New Salisbury had expanded into several days a week, which was detrimental to its own market and others of the region, including that at Old Sarum. On the other hand, complaints about markets that were impoverishing the merchants of Old Sarum by exacting tolls from them, despite their chartered exemption, shows them still trading not only in the region, at places like Lacock, Heytesbury, Chisbury, and Corsham, but in neighbouring counties at Wareham, Blandford Forum, Cranbourne, Christchurch, Middleton, and Southampton, and as far afield as Oxford and Bristol.

Old Sarum still received recognition as a city or a borough, but was increasingly perceived as an appurtenance of the castle, whose keeper was its immediate overlord. The castle continued to be maintained, primarily as an administrative base and prison, and was for some time usually in the hands of the sheriff, but from the late fourteenth century held by lesser custodians and falling into disrepair; in 1515 permission was given to pull it down and make other use of its stones. The borough still was not totally abandoned, but pursued its slow decline into insignificance, all above-ground traces of habitation eventually disappearing.