Constructivism has been adopted by so many opposing factions that a consensus definition seems almost mundane: to acquire knowledge, you must actively construct it through experience, interaction, and reflection; you can’t receive it passively. So to teach something, you can’t just tell—you have to facilitate that active construction.
This is a pretty reasonable claim, given what we know about long-term memory, schema acquisition, elaboration, depth of processing, and so on. It’s clearly not exactly true as stated, given that simple priming effects do exist. Simple exposure to a word list will increase your likelihood of recognizing those words during Tachistoscope presentation, so in some sense it must be possible for you to passively absorb information (though probably attention is necessary for this effect to occur). Anyway, one might regard that as a triviality; for the kinds of knowledge we care about, the broad constructivist claim seems to be roughly true, and certainly underappreciated in almost all educational settings.
People who emphasize constructivism often (over-?)emphasize Inquiry-based learning, discovery learning, problem-based learning, project-based learning, and other styles of learning which involve minimal guidance and explicit instruction. Instead, they tend to promote authentic interest, social embeddedness, complex naturalistic situations, and so on.
As these emphases become stronger, we move into “radical constructivism”, which strongly opposes decontextualization, decomposition, and generally any kind of guided learning activity. RC advocates may claim that all knowledge is subjective, contextual, provisional, and/or interrelational; and that there can be no authorities. Radical constructivists usually insist that all learning be Inquiry-based learning and Situated learning. It’s hard to reconcile these stronger views with the empirical realities of Cognitivism.
Does it work?
Also: How can I reconcile my views on inquiry learning and memory systems?