Sure.  

Why do we have a consensus layer in the first place?  It is a way for us to agree on what transactions are valid and what transactions are invalid.  For example, we all agree that Alice shouldn't be able to move Bob's coins without a valid signature, and that Bob shouldn't be able to create coins out of thin air.  The consensus layer is about obvious stuff like that.  In order for Bitcoin to function as sound money, we need to agree on "black-or-white" rules like this that define which transactions are valid and which are invalid.

Notice that the paragraph above discusses valid and invalid transactions.  No where did I say anything about blocks.  That's because we only really care about transactions in the first place!  In fact, how can a block be invalid just because it includes one too many valid transactions?  

Satoshi added the 1 MB limit as an anti-spam measure to deal with certain limitations of Bitcoin's transport layer--not as a new rule for what constitutes a valid transaction.  We should thus think of every block that is exclusively composed of valid transactions as itself valid.  The size of the block alone should not make it invalid.  Instead, if a block is too big, think of it as likely to be orphaned (a "gray" rule) rather than as invalid (a black-or-white rule).  Perhaps above a certain block size, we're even 100% sure that a block will be orphaned; still we should view it as a valid block!  It will be orphaned because the transport layer was insufficient to transport it across the network--not because there was anything invalid about it.