(a) Critically explain the notion of 'deconstruction' in the light of the postmodern works of Jacques Derrida. 20 marks (b) What…

(a) Deconstruction in the work of Jacques Derrida

Deconstruction is Derrida's way of reading a text closely to show that it rests on hierarchies it cannot hold steady. He set it out in Of Grammatology, Writing and Difference and Speech and Phenomena (all 1967), and later in Dissemination (1972). Derrida insisted that it is neither destruction nor a method to be applied mechanically. It is a reading strategy that shows how a text already undoes its own claims.

Key ideas

• Logocentrism and the metaphysics of presence: since Plato, Western thought has looked for a fixed centre of meaning (God, reason, truth, the conscious subject) that is fully "present". Linked to this is phonocentrism, which treats speech as close to thought and writing as a secondary, even dangerous, copy.
• Binary oppositions: speech/writing, nature/culture, presence/absence, male/female, West/other. One term is always privileged. Deconstruction first reverses the hierarchy and then displaces it, showing that the privileged term depends on the one it excludes.
• Différance: a coined word joining "to differ" and "to defer". Extending Saussure, meaning arises only from differences between signs and is never finally present; it is endlessly deferred along chains of signifiers.
• Trace: every sign carries traces of the other signs it is not, so there is no pure origin of meaning.
• Supplement: Rousseau calls writing a mere addition to speech, but a supplement both adds to and replaces what it supplements, revealing a lack at the "origin".
• **"There is nothing outside the text" (il n'y a pas de hors-texte):** meaning is always mediated by language and context. This is not a denial that the world exists.
• Aporia: points of undecidability where a text's own logic cannot be resolved.

Relation to structuralism and postmodernism. In "Structure, Sign and Play" (1966) Derrida argued that structuralism, including Lévi-Strauss's, still assumed a centre that organises the structure while standing outside it. This made him a key post-structuralist. His attack on fixed centres fits Lyotard's definition of the postmodern as "incredulity towards metanarratives" (The Postmodern Condition, 1979), that is, distrust of grand stories of progress, reason or emancipation. Derrida himself disliked being labelled "postmodern".

Application in anthropology. In Of Grammatology Derrida read Lévi-Strauss's "A Writing Lesson" (Tristes Tropiques), where a Nambikwara chief imitates writing. Lévi-Strauss presented writing as bringing domination into an innocent, speech-based society. Derrida replied that this repeats Rousseau's myth of innocent speech, since the Nambikwara already had forms of "arche-writing" such as the ban on uttering proper names, classification and social violence. Such critiques fed into Writing Culture (Clifford and Marcus, 1986), which treated ethnographies as texts, questioned the ethnographer's authority, and encouraged reflexive, dialogic and multi-vocal ethnography.

Critical evaluation

• Strengths: it exposes hidden hierarchies (colonial, gendered, Eurocentric), supports feminist and postcolonial critique, and made anthropologists aware of the politics of writing.
• Relativism: endless deferral makes it hard to prefer one account over another, and critics see a slide towards relativism or nihilism.
• Habermas and Searle: Habermas held that it blurs philosophy and literature and abandons reason's critical role; Searle challenged Derrida's reading of Austin in their 1977 exchange.
• Politics and practice: textual critique can leave material inequality untouched, and Marxist and practice-oriented anthropologists argue that poverty, violence and bodies are not merely texts. Its obscure style also limits its use in fieldwork.
• Self-reference: its own arguments rely on the logic and clarity it questions.
• Later ethics: Derrida's later work on justice, hospitality and Specters of Marx (1993) shows an ethical and political engagement that answers some of these charges.

Deconstruction is most useful to anthropology as a habit of vigilance about categories and representation, not as a denial that ethnographic knowledge is possible.

(b) Multifactorial traits

A multifactorial trait is one produced by many genes, each with a small additive effect (polygenic inheritance), acting together with environmental factors and their interaction. It differs from a Mendelian trait in the following ways:

• Mendelian (single-gene) traits: one gene of large effect, distinct classes and predictable ratios such as 3:1. Examples are the ABO blood groups, sickle-cell anaemia, albinism and, largely, lactase persistence (a regulatory variant near the LCT gene).
• Multifactorial traits: continuous variation that forms a bell-shaped (normal) curve, and clustering in families without Mendelian ratios. Risk falls sharply as relatives become more distant.

Continuous (quantitative) traits

• Stature: heritability is about 80% in well-fed populations, and genome-wide studies have linked thousands of variants to height. Environment still matters. Heights rose sharply in Japan and Europe as nutrition improved, and in India 35.5% of children under five are stunted (NFHS-5), which shows poor nutrition and infection holding growth below genetic potential.
• Skin colour: controlled by several pigmentation genes (SLC24A5, SLC45A2, OCA2, MC1R and others), with tanning adding an environmental response to UV light.
• Other examples: blood pressure, body-mass index, and the total finger-ridge count, a classic dermatoglyphic trait with heritability near 0.9.

Threshold (discontinuous) traits. For conditions that are either present or absent, Falconer's liability-threshold model assumes that the underlying liability, the sum of genetic and environmental risk, is normally distributed. A person is affected only when their liability crosses a threshold.

Examples:

• Cleft lip with or without cleft palate.
• Neural tube defects: the risk falls with maternal folic acid, an environmental factor.
• Congenital pyloric stenosis: it is commoner in boys, yet relatives of an affected girl face a higher risk because she must have carried more liability (the Carter effect).
• Type 2 diabetes: susceptibility genes such as TCF7L2 combine with diet, physical inactivity and central obesity. Indians tend to develop it at a lower body-mass index than Europeans, and ICMR-INDIAB (2023) put national prevalence at 11.4%.
• Other examples: coronary artery disease, hypertension and schizophrenia.

How such traits are studied

• Heritability (h²): the share of variation in a trait that is due to genetic differences within a given population and environment. It is not a fixed property of the trait.
• Twin and adoption studies: identical twins agree on the trait more often than fraternal twins, which shows a genetic role. But identical twins do not agree 100% of the time, so the environment also plays a part.
• Recurrence risk: it is higher when more relatives are affected, when the case is more severe, or when the affected person belongs to the sex that is less often affected.
• GWAS and polygenic risk scores: these test the "common disease, common variant" hypothesis, in which many common variants of small effect add up to the risk.

(c) Sampling techniques in selecting the study group

The population (universe) is every unit relevant to the research. The sampling frame is the list from which units are drawn, such as a voter list or a village household register. The sample, or study group, is the subset actually studied. A census (complete enumeration) is practical for a small community, such as a single village or a small tribe like the Toda. Large populations need a sample, which saves time and cost and, if drawn properly, still supports generalisation.

Probability sampling gives every unit a known, non-zero chance of selection, so the results can be generalised with a measurable sampling error.

• Simple random sampling: units are drawn by lottery or random numbers, for example households from a village register. It is unbiased but needs a complete frame.
• Systematic sampling: every *k*th unit is taken after a random start. It is easy in house-to-house surveys but is biased if the list has a hidden periodic order.
• Stratified sampling: the population is divided into homogeneous strata (tribe, caste, sex, age, landholding) and each stratum is sampled proportionately or deliberately over-sampled. It is well suited to heterogeneous Indian settings, for instance comparing nutrition among Scheduled Tribe, Scheduled Caste and other households.
• Cluster sampling: whole groups (villages, hamlets, schools) are selected at random. It is cheap for scattered populations such as hill tribes, but units within a cluster resemble each other, which increases error.
• Multi-stage sampling: stages are combined. The National Family Health Survey, for example, first selects villages or urban blocks and then households within them.

Non-probability sampling suits intensive ethnographic work, where depth matters more than statistical generalisation.

• Purposive (judgement) sampling: choosing key informants such as elders, shamans or midwives, or a "typical" village.
• Snowball sampling: reaching hidden or stigmatised groups (drug users, sex workers, transgender communities, informal migrant workers) through referrals.
• Quota sampling: filling fixed numbers per category, such as age or sex, without random selection.
• Convenience sampling: whoever is available at a market or clinic. It is quick for pilot studies but highly biased.
• Theoretical sampling: in grounded-theory work, selecting new cases until no new themes emerge (saturation).

Choosing the right technique

• Research aim: estimating prevalence (anaemia, stature, literacy) needs probability samples; understanding rituals, kinship or meanings calls for purposive selection.
• Nature of the population: a small, bounded community allows a census; large, diverse populations need stratified or multi-stage designs; scattered settlements favour cluster sampling.
• Availability of a frame: nomadic or remote groups often have no lists, so fieldworkers first map and enumerate households, the classic "village census" at the start of fieldwork.
• Sample size and error: sampling error falls roughly with the square root of sample size. For a proportion near 50%, about 385 respondents give a ±5% margin at 95% confidence.
• Bias and ethics: non-response, gatekeepers steering the researcher to favoured informants, seasonal migration, and the need for informed consent all shape the final study group.

In practice, anthropologists often combine a probability survey for breadth with purposive in-depth cases for meaning.