I have discussed this admirable acount by John Mepham, once a student of Hall's at the Birmingham Centre, in my own (1992) work. Here is an extract:
 
 

... Mepham offers a very clear way of rejecting the earlier work on consciousness in Marx and Marx and Engels, especially in the German Ideology (Marx and Engels 1974). This work saw consciousness and material reality as separate entities, and offered terms like 'inversion' to explain the ideological relation between them, as in the famous 'camera obscura' metaphor. In brief, ideology in this early work takes the form of 'false consciousness', inverted reality, 'illusions' or 'phantoms'. 

Rather as in the Heck piece summarised [here actually] , Mepham reads ideology as a matter of discourses. He characterises the new model, discoverable in Capital, as a 'serial' one, rather than a model of 'base/superstructure',  with categories like 'real relations', 'phenomenal relations', 'ideological categories' and 'discourse practices' connected by structured practices - 'transformation', 'representation', 'generation', and 'reproduction'. 

Reality in this model is not simply inverted or mystified but 'misrecognised' as an effect of discourses which, to be brief, work on the 'phenomenal' level, deal with deceptively simple surface forms, and maintain an account which leaves out the complexities in the depth. Mepham's analysis is developed in terms of 'common sense' or 'spontaneous' thought, and formal ideas - political economy and philosophy - which incorporate the distortions of everyday life and make a science or a system out of them. In both science and common sense, dubiously phenomenal categories support each other to create a whole 'semantic field'. However, this self-sustaining field is subject to disruption from the very dynamics of capitalism itself which continually corrodes the boundary between the real and the phenomenal as it forces social change.

And in my 1996 book, I discussed him again, this time in terms of the debate in Sociology about 'embedding': 
 

To borrow from a discussion in Mepham and Ruben (1979), for example, embedding takes place:

1. in 'real relations' (e.g. the actual relations between bosses and workers in production),
2. via the 'phenomenal forms' we discussed above, which can scarcely depart from the system that produces them and thus appear as 'natural', 
3. in 'ideological categories' (more developed 'scientific looking' but still rooted in the phenomenal forms at crucial moments, and, Lecourt tells us, they are ultimately nested within practical ideologies -- see Macdonell 1986). 

Whether these processes are ranked in some sort of order of importance is another large area of debate of course, and one which is at the heart of the controversy about determinism in marxism. Ultimately, we have implied so far, all the categories (at least) of belief systems can be traced back to a mode of production that limits, constrains, embeds or determines them in some final sense, despite their apparent or even their real autonomy in some respects. 

To return to the issue of modern culture, though, there are many examples around us these days that Marx did not address directly, of course. Above all, there are problems with those which seem to have emerged from distinct and separate cultural practices, or to have been so worked upon and developed, so transformed by the activities of autonomous specialists, that they have escaped any conclusively determining traces of their ultimate origins.

One response is to add to lists like Mepham’s what he calls 'discourse practice' in order to try to come to grips with these modern examples, and, no doubt to respond specifically to emerging work on the 'society as text' metaphor. It was thought possible to further extend marxist analysis into a wide range of familiar modern cultural practices, granting a certain amount of autonomy to them, but preserving a link to overall systems of domination (as Thompson 1984 puts it). We shall consider a number of alternative positions in marxist linguistics and in ‘post-marxism’ in what follows, but in general the problem is obvious: how much can continue to be added to marxism while insisting that it is still marxism?