Insulin resistance (IR) is an important biochemical phenomenon because it is closely linked to major and highly prevalent diseases including diabetes mellitus, atherogenic dyslipidaemia, the fatty liver disease dysfunction, and ovulatory dysfunction. Yet major barriers to understanding the mechanisms linking IR to these clinical diseases have included 1. The difficulty in discerning cause and effect relationships in associated phenomena in a complex multisystem disorder and 2. The limitation of conventional definitions of IR, which rely solely on impaired insulin action on glucose metabolism. Key analytical approaches to circumventing these problems include the creation of targeted genetic alterations in mice, and the study of naturally occurring single gene disorders of insulin action in humans. These complementary approaches over recent years have led to growing interest in the idea that fatty liver and metabolic dyslipidaemia only appear when IR is partial, affecting the hypoinsulinaemic actions of insulin without impeding insulins stimulatory effects on hepatic de novo lipogenesis. This means that compensatory hyperinsulinaemia can act through unaffected signalling pathways to drive liver fat accumulation and hypersecretion of hepatic VLDL. I shall critically review the evidence for this concept, focusing on metabolic studies in humans with rare single gene defects associated with severe IR.