The physical nature of the apparently densest groups of galaxies, known as compact groups is a topic of some recent controversy, despite the detailed observations of a well-defined catalog of 100 isolated compact groups compiled by Hickson (1982). Whereas many authors have espoused the view that compact groups are bound systems, typically as dense as they appear in projection on the sky (e.g., Williams & Rood 1987; Sulentic 1987; Hickson & Rood 1988), others see them as the result of chance configurations within larger systems, either in 1D (chance alignments: Mamon 1986; Walke & Mamon 1989), or in 3D (transient cores: Rose 1979). As outlined in the companion review to this contribution (Mamon, in these proceedings), the implication of Hickson's compact groups (HCGs) being dense bound systems is that they would then constitute the densest isolated systems of galaxies in the Universe and the privileged site for galaxy interactions. In a previous paper (Mamon 1986), the author reviewed the arguments given for the different theories of compact groups. Since then, a dozen papers have been published on the subject, including a thorough and perceptive review by White (1990), thus more than doubling the amount written on the subject. Here, the author first enumerates the arguments that he brought up in 1986 substantiating the chance alignment hypothesis, then he reviews the current status of the numerous recent arguments arguing against chance alignments and/or for the bound dense group hypothesis (both for the majority of HCGs but not all of them), and finally he reconsiders each one of these anti-chance alignment arguments and shows that, rather than being discredited, the chance alignment hypothesis remains a fully consistent explanation for the nature of compact groups.