The diagnosis and classification of pelvic girdle pain (PGP) disorders remains controversial despite a proliferation of research into this field. The majority of PGP disorders have no identified pathoanatomical basis leaving a management vacuum. Diagnostic and treatment paradigms for PGP disorders exist although many of these approaches have limited validity and are uni-dimensional (i.e. biomechanical) in nature. Furthermore single approaches for the management of PGP fail to benefit all. This highlights the possibility that 'non-specific' PGP disorders are represented by a number of sub-groups with different underlying pain mechanisms rather than a single entity. This paper examines the current knowledge and challenges some of the common beliefs regarding the sacroiliac joints and pelvic function. A hypothetical 'mechanism based' classification system for PGP, based within a biopsychosocial framework is proposed. This has developed from a synthesis of the current evidence combined with the clinical observations of the authors. It recognises the presence of both specific and non-specific musculoskeletal PGP disorders. It acknowledges the complex and multifactorial nature of chronic PGP disorders and the potential of both the peripheral and central nervous system to promote and modulate pain. It is proposed that there is a large group of predominantly peripherally mediated PGP disorders which are associated with either 'reduced' or 'excessive' force closure of the pelvis, resulting in abnormal stresses on pain sensitive pelvic structures. It acknowledges that the interaction of psychosocial factors (such as passive coping strategies, faulty beliefs, anxiety and depression) in these pain disorders has the potential to promote pain and disability. It also acknowledges the complex interaction that hormonal factors may play in these pain disorders. This classification model is flexible and helps guide appropriate management of these disorders within a biopsychosocial framework. While the validity of this approach is emerging, further research is required.