Early Celtic Social Structures



One of the few sources we have about ancient Celtic society on the continent from this period is the text by Julius Cæsar during his time in Gaul 58-50 BCE. How accurate a view does he provide of Gaulish society? Cæsar wrote for his fellow Roman nobles and uses Roman words (such as equites “horse-mounted warriors”) to describe Gaulish society. It is very likely that he imposed Roman concepts and principles on Celtic ones in order to make Gaulish society look familiar to the people of Rome and be easily understood by them. Thus, Cæsar’s text probably makes Gaulish society look more Roman society than it actually was. Some of the terms he gives, however, are specifically Gaulish in origin and are unlike aspects of Roman society, which adds weight to their veracity.

The kin-group – a set of people sharing some familial relationship, whether by descent or contract (such as marriage or adoption) – is central in the operation of virtually every society. Although the kin-group may be defined differently in different cultures, it is usually by reference to the kin-group that property is inherited, legal rights are recognized, and privileges are asserted.

Celtic societies were largely rural in character in this era, consisting of clusters of houses scattered across the countryside; they produced over 90% of the resources that they consumed on their own land. As was typical in this period, the Celts practiced partible inheritance, meaning that after death, the wealth (including land) of a father was split equally amongst all of his sons. Thus, the general pattern was for men to set up their homes on the territory of their paternal kin-group and for their wives to move to their territory. The neighbours of most men, therefore, were related to them through their fathers and their land could be annexed if they died without heirs. The pressure of the growth of families, the constant division of property and environmental conditions (such as soil depletion) could force people off their holdings and seek a livelihood in other communities under very compromised conditions.

Ancient Celtic society can be reconstructed conceptually on the basis of terminology: by comparing words attested in Gaulish, Old Gaelic and Old Welsh, a number of common principles and structures that seem to underly Celtic society and guide its development from small-scale kin-groups to larger-scale nations has been identified. One of the main principles inherent in the system is that of the stature of the senior male of any line of descent: once a man became the oldest surviving member of his lineage, he held its land on its behalf and was no longer dependent or subservient to any other member of it legally or socially; he was effectively responsible for running its affairs, enforcing law within it, and negotiating on its behalf with other senior males. Lesser-ranked males – cousins descended from the same ancestor, his sons, and so forth – were either dependents within his own household or else attached as vassals to other households.

Independent adult males who have inherited and become the heads of households are referred to in Proto-Celtic as *wiros (Gaulish uiros, Old Gaelic fer, Old Welsh gŵr), and are in distinction to the junior men or vassals who are dependent upon them, Gaulish mapon os (or uasseli tus), Old Gaelic mac (or foss), Old Welsh mab (or gwas). The Gaulish term uasseli (for servant or dependent) was borrowed into Latin and thence into Germanic languages, emerging in feudal terminology as vassal (and French valet).

The term for kin-group in Proto-Celtic is *kenetlom (Old Gaelic cenél, Old Welsh cenedl). Caesar uses the Latin term pagus (pl. pagi) to refer to some primary social formation, although it is unclear whether it is the extended kin-group (a coalition of extended families, like a clan) or some larger political entity. A pagus could have been led by a dominant head of kin-group who took in other kin-groups as dependent vassals on the basis of his personal charisma, command of resources, political expertise and military muscle.

Cæsar seems to indicate that two or more pagi formed confederations which he refers to with the Roman term civitas. This is translated in this sourcebook as “nation.” (In older texts, the term “tribe” is often used.) Caesar indicates that in some cases, a civitas had its own magistrates, who are referred to by the Gaulish term vergobret. This term seems to be made of the Gaulish words *vergo “work” and *brito “judgment,” suggesting something like a high judge or commissar. This is suggestive of how Gaulish nations were moving away from kin-based leadership to the formal institutions of a state. Excavations in 1978 in central France uncovered an inscription on pottery mentioning a vergobret.

The Celtic legal and economic system seems to have intentionally fostered the middle ground for households between dependency and independency: the household was encouraged to be as self-sufficient as possible but fell back on the wider kin-group as an insurance policy. Households could draw from the common resources of the kin-group when an individual or household did not have sufficient means, such as when paying for crimes. On the other hand, the kin-group could intervene when a member’s business put others in the kin-group at risk.


Slavery is common in societies that practice partible inheritance because slaves provide labour without competing with siblings for inherited land which would be further divided. Slaves tended to be captured as children or teenagers by professional raiders or slavers and sold into homes of practically all social ranks.

Because they lived and worked alongside their owners who were not responsible for the violence of their capture (and the gruesome conditions associated with it), slaves could generally form strong bonds with them. Those who survived into middle age could be freed from their servitude and be granted land on the property of their former owners, and their offspring were enfranchised by the kin-group. The Proto-Celtic word for slave is *magus. Slaves were not physically or visibly different from anyone else in the community (race-based slavery did not emerge until the 17th century).

Saint Patrick, a Briton alive during Roman occupation, was enslaved by Irish pirates to tend flocks in Ireland. In his Epistle he denounces the practice of slavery by Insular Celts, especially when Christians were enslaved by people who were supposed to have been fellow Christians.


One of the important social institutions of Celtic societies that was not governed by genetic relationship was fosterage. Much more detail exists about fosterage in Goidelic society survives than in Brythonic society, but they probably did not differ in the main. It was considered a great privilege to be the foster-parent of a child, usually given by parents of a higher status to those of a lower status. Fosterage was arranged by contract and there were three discrete phases of life during which fosterage could happen: birth to seven years; seven years to twelve; twelve years to seventeen. The child was accompanied by a fee (typically in cattle) which was returned when the contract of fosterage was complete.

Fosterage allowed families to form strong bonds across kin-groups; in fact, the bond of fosterage is generally presented as being as strong as that of kinship. It also allowed the child to learn valuable skills from specialists outside the immediate family (a fee was imposed if the educational aims were not met). Another rationale, however, was to keep children safe from the harm that could be inflicted on them by jealous rivals who were competing with them for the inheritance of power and wealth and could eliminate them by murder or maiming. Such conflict was bound to happen in families of children with more than one mother (i.e., wealthy, freemen could have more than one wife and hence many eligible children).

Kings, Kingdoms and Nations

Different parts of Gauls were developing different models of governance. By the time that Cæser went to Gaul there was a mixture of inherited and elected rulers and governing councils in various nations.

The Celtic words for “king” – Gallo-Brythonic rīx and Old Gaelic – are cognates of the Latin rex, and all of these derive from the Indo-European verbal root *Hreg, meaning “to stretch out straight.” The English word “regulate” (from a Latin borrowing) conveys a similar sense. Thus, early kings were intended to keep ordered, regulated, and strong the community whom they represented, who had invested their trust and interests in his person. The literary sources often emphasize the sacred and symbolic aspects of kingship, while the historical texts often emphasize the actions and political exercise of power. Celtic kingship is thus a multifaceted institution which was shaped by many factors; different kinds of evidence highlight different aspects of kingly power.

The ideology of Celtic kingship held that it was a sacred institution in which the king, as representative of his human community, was married to the sovereignty goddess of the land (a motif common to many cultures). The inaugural ceremony itself reflected this idea. His good rule ensured prosperity for his people, while poor rule would sour his relationship with the divine, spoil the terrestrial condition of his people and ultimately bring about his downfall.

The Proto-Celtic term *toutā (Celt-Iberian touta, Gaulish touta, Old Gaelic túath, Old Welsh tud) referred to the people ruled by a king: a kingdom. This Celtic term may correspond to Cæsar’s “civitas.” It has a close Germanic cognate which gives the modern name for the German language (Deutsch). Celtic nations probably started as extended kin-groups but eventually grew in extent and formality, grafting dynasties together by claiming common descent from famous founder figures, deities or assumed qualities, as sometimes reflected in the names of early Celtic nations.

According to Caesar, in some parts of Gaul the civitas was ruled by a senate, in others by a ruler (such as a king), and sometimes by some combination of the two. Caesar also mentions a Gaulish council with broader territorial influences made up of prominent members of the senates of several nations and other leading men of Gaul, probably an international body meant to resolve disputes.


In principle, Celtic kingdoms/nations were self-contained units whose integrity was not to be impinged upon by others; in other words, the laws and rule of a nation and their leader did not extend beyond their own boundaries. We can see this reflected, for example, in Caesar (§ 7.33): “by the laws of the Aedui, it was not permitted those who held the supreme authority to leave their territory […].” Thus, the king (and his kingdom) corresponds closely to the notion of the head of the kin-group in that both men are responsible for managing affairs within the bounds of their population, enforcing laws internally and negotiating external affairs with their peers who ruled other groups.

Class Discussion

What are the differences between the ways in which Celtic kings in different nations gained their office and exercised their power in the following passages?


One of the complexities of law is always how to guarantee that one or both parties will carry out the terms of a contract faithfully. This is often referred to as a “pledge” or “surety.” Many societies of this era practiced giving or exchanging hostages as a means of ensuring contracts between parties, especially subordinate relationships between kin-groups or kingdoms.

Hostages would normally be submitted voluntarily by a vassal group to their overlords but could also be captured after forcing a kin-group or kingdom to submit to an overlord. Hostages were typically the sons of the vassal group, but daughters sometimes served as vassals as well. If the subject group rebelled against the authority of the overlord, hostages could be maimed, blinded, ransomed, or killed.

Irish law tracts make it clear that overlords were expected to hold hostages in their banqueting-hall, no doubt to demonstrate publicly their superiority over subject kings. The pre-historic Irish king Níall Nóigíallach [Old Gaelic “of the Nine Hostages”] was so named because of his collection of hostages from subject kingdoms.

Class Discussion

How do the following passages demonstrate the practice of the holding of hostages? Who is giving up (or asked to give up) hostages to whom, and why?


A ruler needed to have a retinue to help with the duties of his office and to provide him with the military muscle to ensure his safety and enforce his commands. The size of a ruler’s retinue generally corresponded directly to his social status: the greater his status, the more followers he had. This, however, posed challenges for paying and rewarding them; there were always natural limitations to expansion. Members of a retinue would typically be foster-sons or the sons of vassals who had not yet inherited land and seniority from their fathers.

Caesar records the Gaulish word for the followers of the king in his retinue: ambactus, meaning “to act around someone.” This Celtic word was borrowed into Latin, and then into German from Latin, resulting in French ambassade and English “ambassador.”

It was probably the ambitious males who had come of age but had not become head of kin-group, been hired into a retinue or fit into the system of vassalage who sought employment as mercenaries, the Gaesatae of the continent or the fian of Goidelic society.

Class Discussion

Who are the members of the retinues, and their functions, as depicted in the following passages?


Dunham, “Caesar’s perception.”

Fraser, From Caledonia.

Karl, “Neighbourhood.”

—, “Random Coincidences.”

— “The dutch Group.”

Kelly, Early Irish Law.

Koch, Celtic Culture.

Meid, The Celts.

Woolf, From Pictland.