Prehistoric Archaeology I: The Concept and Its Sub-Fields from Ethno-Archaeology to Action Archaeology, the Theoretical Paradigms from the Descriptive to the Scientific and the Interpretative Periods, and the Methods of Dating from Typology and Seriation to Radiocarbon, Potassium-Argon, Thermoluminescence and Palaeomagnetism, with the Radiocarbon and K-Ar Arithmetic Worked

Archaeology is the study of the human past through its material remains — artefacts, features, ecofacts and their contexts; prehistoric archaeology is that study for the immense period before written records, which in India means everything before the Harappan script (itself undeciphered) and, for most of the subcontinent, before the Mauryan inscriptions. Daniel Wilson coined "prehistory" in 1851; Christian Jürgensen Thomsen's Three-Age System (Stone, Bronze, Iron; Copenhagen, 1836) and John Lubbock's division of the Stone Age into Palaeolithic and Neolithic (Pre-historic Times, 1865) gave it a frame, and Robert Bruce Foote's handaxe from Pallavaram near Madras (1863) began it in India. Because there are no texts, prehistoric archaeology is by necessity anthropological: its questions are about subsistence, technology, settlement, social organisation and belief, and its answers come from the same comparative and ecological reasoning as the rest of the discipline. Its sub-fields specialise that reasoning. Ethno-archaeology studies living peoples — their tools, butchery, house-building, pottery, discard — to build analogies for reading the archaeological record (Binford among the Nunamiut, 1978; Gould in Australia; Indian studies of village potters, pastoral camps and hunting-gathering groups). Experimental archaeology reproduces past technologies and processes under controlled conditions — knapping handaxes and Levallois cores, smelting copper, building and burning houses (the Overton Down earthwork, 1960; Coles's Experimental Archaeology, 1979). Environmental archaeology reconstructs past climate, landscape, plants and animals from pollen, seeds, bones, molluscs, sediments and isotopes (archaeobotany, zooarchaeology, geoarchaeology) to place sites in their ecology. Settlement archaeology studies the distribution of sites across a landscape and the layout within them — site catchment analysis (Vita-Finzi and Higgs, 1970), central-place models, Willey's Virú valley survey (1953) and Chang's work — to read economy and social organisation. Cognitive archaeology (Renfrew, 1982; Mithen) seeks past ways of thought — symbolism, religion, planning, measurement — from art, burial, monuments and tool design; the Bhimbetka paintings and the Harappan weights are its Indian material. Geo-archaeology applies geology and geomorphology to sites: stratigraphy, sedimentology, soils, site formation, palaeochannels (the Ghaggar–Hakra), the terraces of the Soan and Narmada. Action archaeology (Kleindienst and Watson, 1956, on the model of Sol Tax's action anthropology) is archaeology done with and for a living community, in its interest — recording the recent past, returning results, protecting sites; it shades into public and community archaeology.

The descriptive period (roughly 1840s–1950s), also called culture-historical archaeology, was concerned with what happened where and when: building chronologies by stratigraphy, typology and seriation, defining "cultures" as recurring assemblages of artefact types (Childe's definition in The Danube in Prehistory, 1929), mapping them and explaining change by migration and diffusion. Its achievements were the Three-Age System, the Palaeolithic sequence of de Mortillet, the Indian sequence of Foote, de Terra and Paterson and Wheeler's stratigraphic excavation; its limitation was that it described sequences without explaining processes. The scientific period, or New Archaeology (1960s–1980s), announced by Lewis Binford's "Archaeology as Anthropology" (1962) and by David Clarke's Analytical Archaeology (1968), demanded explanation: archaeology should be a science, testing hypotheses about cultural process (hence "processual"), using systems theory, ecology, statistics, sampling, ethno-archaeology and "middle-range theory" to link the static record to the dynamic past, and treating culture as an adaptive system. Its tools were the computer, the radiocarbon date, site catchment analysis and the settlement survey; its founding ambition was cultural evolution in Leslie White's and Julian Steward's sense. The interpretative period, or post-processual archaeology (from the 1980s), led by Ian Hodder (Symbols in Action, 1982; Reading the Past, 1986), Shanks and Tilley, and Christopher Tilley, objected that the New Archaeology's science ignored meaning, agency, ideology and the archaeologist's own position: material culture is "meaningfully constituted", to be read like a text; the individual acts and is not merely adapted; interpretation is always from a standpoint (gender, class, nation), so the discipline must be reflexive and admit multiple readings, including those of indigenous and descendant communities. Hodder's own excavation at Çatalhöyük (from 1993) was designed as a reflexive project. The three periods succeed one another as emphases, not as replacements: chronology building, processual explanation and interpretative reading are all in use, and Indian archaeology, still largely engaged in building sequences, has taken up processual settlement studies (the Bhima, Krishna and Ghaggar surveys) and, more recently, interpretative work on Harappan identity and rock art.

Relative dating orders things — earlier, later, contemporary — without giving years; absolute (chronometric) dating gives an age in calendar or radiometric years with an error term. The oldest relative method is stratigraphy itself: Steno's law of superposition (1669) — in undisturbed deposits the lower layer is the older — which Wheeler's grid and section-drawing brought to Indian excavation. Typology arranges artefacts in series by form on the assumption that types change gradually and directionally, so that a sequence can be read from the series (Montelius, 1903, on Bronze Age axes; de Mortillet's Palaeolithic stages; Bordes's Mousterian types); it dates by association, and it can mislead when styles revert or several forms coexist. Seriation orders assemblages, not single objects, by the overlap of types: Flinders Petrie's sequence dating of the predynastic Egyptian graves at Naqada (1899) arranged pottery so that each type's frequency rose and fell smoothly — the "battleship curve" of frequency seriation later formalised by Ford, Brainerd and Robinson (1951) — while contextual seriation orders by presence and absence. Geo-archaeological dating uses the geological record: river terraces (the Soan sequence of de Terra and Paterson, 1939, whose terraces were matched to Himalayan glacial and interglacial stages), glacial moraines, loess, sea-level stands and raised beaches, correlated with the Pleistocene climatic sequence (next chapter), so that an industry on a terrace is dated by the terrace. Chemical dating of bone rests on changes after burial: fluorine from groundwater is absorbed into bone at a rate depending on the local water, uranium is likewise absorbed, and nitrogen (from collagen) is lost, so that bones from the same deposit can be tested for contemporaneity by their fluorine–uranium–nitrogen (FUN) content — the method by which Kenneth Oakley exposed the Piltdown forgery in 1953, the "fossil" jaw having far less fluorine than the deposit's genuine bones; it is a relative test within one site, since rates differ between sites. Cross-dating transfers a known date to an undated context through associated objects — a Roman coin in an Indian megalith, Harappan seals in Mesopotamian levels of Sargonic date (about 2300 BCE), Mesopotamian objects at Harappan sites — or through a dated stratigraphic sequence matched from one site to another; it was the backbone of Old World chronology before radiocarbon and remains the way protohistoric India is tied to Mesopotamia and the Persian Gulf. Obsidian hydration and amino-acid racemisation, though listed with the absolute methods, are in practice calibrated relative methods and are treated in the next section.

Radiocarbon dating, devised by Willard Libby at Chicago in 1949 (Nobel Prize in Chemistry, 1960), rests on the radioactive isotope carbon-14, formed in the upper atmosphere by cosmic-ray neutrons striking nitrogen-14, oxidised to CO₂ and taken up by every living plant and, through the food chain, every animal, so that living tissue holds a constant proportion of ¹⁴C to ordinary ¹²C. At death intake stops and the ¹⁴C decays back to nitrogen at a fixed rate: Libby measured the half-life as 5,568 years (the "Libby half-life", still used by convention in reporting uncalibrated dates), and the Cambridge value of 5,730 ± 40 years is the accepted physical figure. Measuring the ¹⁴C remaining — by counting beta decays (conventional method, needing grams of carbon) or, since the late 1970s, by counting the atoms themselves in an accelerator mass spectrometer (AMS, needing milligrams and extending the range) — gives the time elapsed. The fraction remaining after n half-lives is (1/2)ⁿ: after one half-life 50%, after two 25%, after three 12.5%, after four 6.25%, after ten about 0.1% — the practical limit is about 50,000 years (roughly nine half-lives), beyond which too little ¹⁴C survives to measure. Worked: a charcoal sample retains 25% of the ¹⁴C of a living plant; 25% = (1/2)², so two half-lives have elapsed and the age is 2 × 5,730 = 11,460 years (2 × 5,568 = 11,136 Libby years). A bone with 12.5% remaining is three half-lives, 17,190 years, old; one with 6.25% is four, 22,920 years. For a fraction that is not a power of one-half the formula is t = (5,730 ÷ 0.693) × ln(N₀/N) = 8,267 × ln(N₀/N): a sample with 70% remaining is 8,267 × ln(1.4286) = 8,267 × 0.3567 ≈ 2,950 years old. Dates are reported as years BP (before present, "present" being 1950) with a ± one-sigma error, and because the atmospheric ¹⁴C level has varied, raw dates are calibrated against tree rings, corals and varves to give calendar years (cal BP or cal BCE); materials are charcoal, wood, bone collagen, shell, peat and seeds; contamination by younger carbon (roots, humic acids) and the reservoir effect in marine and freshwater samples are the sources of error. Potassium-argon dating uses the decay of the radioactive isotope potassium-40, which makes up about 0.012% of all potassium, to argon-40 (about 11% of decays; the rest go to calcium-40) with a half-life of about 1.25 (1.248) billion years. When lava or volcanic ash cools, any argon escapes and the clock starts at zero; thereafter ⁴⁰Ar accumulates in the crystal lattice, and the ratio of ⁴⁰Ar to ⁴⁰K measured by mass spectrometry gives the time since crystallisation. Because the half-life is so long, the method dates only material older than about 100,000 years (below which too little argon has formed), but it has no upper limit and so covers the whole of hominin evolution: it dated Olduvai Bed I to 1.75 million years (Leakey, Evernden and Curtis, 1961), the Hadar and Koobi Fora tuffs, and the Java hominin beds. The refined ⁴⁰Ar/³⁹Ar method (irradiating the sample so that ³⁹K becomes ³⁹Ar, then measuring both argon isotopes in single crystals) removes the error from unrelated potassium and from inherited argon and is now standard — it produced the Omo I dates of 2005 and 2022. K-Ar dates the volcanic layer, not the fossil: a hominin between an underlying tuff of 3.4 million years and an overlying one of 3.2 million is bracketed, not dated directly.