The Three-Age System was embraced in Scotland before England (Rowley-Conwy 2007 and see also the ScARF Neolithic Panel report). In the first synopsis of the Scottish Iron Age, Joseph Anderson (1883) insisted that this period should not be assigned absolute dates as he felt, understandably at the time, that prehistory could have no specific chronology. Since then, the term ‘Iron Age’ has been used in Scotland for a period beginning as late as the first century BC, a full four centuries after it was understood by Hawkes and Kendrick (1931) to begin in southern England on the basis of theories involving “iron-using, Celtic-speaking colonists” spreading slowly up-country (Piggott 1958, 75). This diffusionist perspective (together with its exaggerated time-lag), however, became unsustainable in the face of new evidence and Piggott (1966, 3) subsequently backdated the inception of iron-using to 550BC. Following the radiocarbon revolution (Renfrew 1973), later writers moved the date even earlier, to the seventh-eigth century BC (e.g. Harding 1974, 14; Ritchie and Ritchie 1981, 89). The Iron Age has in the past been understood to terminate with the Roman invasion of AD78 or in the third century AD or later if Piggott’s (1966, 3) scheme is followed. Despite problems discussed in more detail below, this is still used by some researchers (e.g. Armit 1997; Armit and Ralston 2003), sometimes in a modified form (e.g. Hingley 1992, in which the terminal date was set at AD200, in order to separate clearly the Picts as an early Medieval phenomenon).
However, a project by Needham et al. (1997), aiming to establish an independent chronology for British Bronze Age metalwork through a programme of radiocarbon dating of associated organic materials, has led to a revision of the dating of LBA metalwork assemblages. This includes the backdating of the end of Ewart Park metalwork from 700BC to c. 800BC, suggesting that the LBA-IA transition should also be backdated by about a century (Needham 2007; but cf O’Connor 2007). There is no good reason in the evidence to suggest a time-lag between the development of styles of metalwork in later prehistory in different areas of Britain. There is a danger, of course, that the dating of events in the Early Iron Age is propelled backwards as a result of the ‘plateau’ in the radiocarbon calibration curve, which begins at around 800BC. Such uncertainties can only be resolved through future work, and for now it would seem reasonable to use 800BC as a useful marker for the beginning of the Iron Age (though the question of the introduction of iron is another contentious one; there is some evidence for its use in Britain from the 10th century BC (Collard et al. 2006), but very little sign of its early use in Scotland. The use of Bayesian statistics to separate out the AMS dates that fall within the plateau on the radiocarbon calibration curve is a highly promising avenue for further research.
Since the late 1970s/early 1980s, however, some archaeologists working in Scotland have adopted a chronological scheme known as the long Iron Age based on a Scandinavian model, breaking down perceived barriers between the Late Bronze Age, the Iron Age and the Early Historic period and taking incursions by the Norse, rather than the Roman army, as the terminus of the period (e.g. Chapman and Mytum 1983; Ralston 1980; Haselgrove et al. 2001, 3; Harding 2004; Haselgrove et al. 2001, 3). The long Iron Age therefore covers the first millennia BC and AD (as epitomised in the naming of the First Millennia Studies Group), and has been summarised by Parker Pearson and Sharples (1999) thus:
|Early Iron Age||700-100BC|
|Middle Iron Age||200BC-AD400|
|Late Iron Age||AD 300-900|
A somewhat looser definition of the Later Iron Age, “in the first millennium AD prior to the Norse settlement” (Armit 1990b) was adopted for the Scottish Archaeological Forum of 1988; as Armit (1990b, 1-2) explained “…no precise dates [were] given for this…the division [was] clearly an arbitrary one…”. Since then, the term Late/Later Iron Age has been used more precisely. Sharples and Parker Pearson (1999) define it variously as AD300-900 or AD400-800. Downes and Ritchie (2003) quote it as AD300-800. Most recently, Harding (2004, 3) has argued for a “‘long’ Iron Age, in which ‘early’ represents a span of time that in Southern Britain would cover the whole of the pre-Roman Iron Age, and ‘late’ is applied to the first millennium AD from around its second quarter”. The 250 years between these two brackets becomes the Roman Iron Age. Harding rejects the use of the term Middle Iron Age in a Scottish context, arguing that it constrains the occupation of brochs to “a limited span of two or three centuries around the turn of the millennium”. This perceived failing would seem to be no more than conventional depending upon a rigid classification of brochs as ‘middle Iron Age’.
The usefulness of the term ‘long Iron Age’ has become particularly apparent in the Western and Northern Isles (e.g. Armit 1990a; Downes and Ritchie 2003), where in the past cellular buildings have been described as ‘Pictish’, despite the evidence for local continuity in architectural tradition, the lack of written records and the geographical distance from the Pictish heartland. The greater chronological fluidity offered by the term allows the archaeologist to appreciate the evolution of architectural traditions and social development in the longue dureé and over wide, and environmental very distinct, areas. The use of this long Iron Age reflects a distinct movement of interest towards the study of the Northern and Western Isles over the past twenty years. In the south and east of Scotland, the Norse incursion can be no more a logical stopping point than the Roman invasion or the first historical mention of the Picts is in the Atlantic north and west. There is much to be said for the adoption of a chronological scheme that does not lay so much stress upon the significance of the dates of the earliest surviving written records, and the rather artificial divisions between later prehistoric, Roman and Early Historic periods which result from this.
It is clear from this discussion that there is no universally accepted chronological scheme for the Scottish Iron Age. Although it has been said that the traditionally defined Iron Age (quoted as 700BC to AD500) merges ‘imperceptibly’ into the Early Historic (Armit 1997c, 15), there is evidence for major changes in the settlement record of many areas from c. AD 400 if not from c. AD200.