In any therapy, assumptions must be made. Cognitive-behavior therapy would assume cognitive distortions. Person-centered therapy would assume some deeper meaning (technique is scarce with the theory . . . the theory itself can use multiple). Psychodynamics uses interpretation, which is an assumption about innate, subconscious, preconscious, or unconscious motives that the client most likely cannot understand or identify.
Assumptions, though, are inherently tricky. The clinician has data from different sources: the client, the clinician's own ideas, and the clinician's study. Two of these three sources do not come from the client; they come, instead, from books (experiences of other counselors) and the clinician's own experiences. Neither of these can ever be 100% accurate, as neither have the knowledge about the client that the client does. So, it is important to use the client as a consultant to themselves.
Blundering through an interview can hurt the relationship and set back any positive change. A blunder, in this case, is meant as an interpretation that is either incorrect and offensive to the client or correct and improper at the time. The former occurs when the client knows the interpretation's matter and the clinician does not. In this case, the client is not consulted on the meaning of their behaviors and is offended by the clinician's view of them. The latter can occur when the interpretation breaks previously indicated (or perhaps not) norms in past sessions.
So, before any assumptions are made in a session, the client must be consulted. Simple questions, such as : What do you think here? Or an even more personal: I'm wondering what you're thinking about this. More data is always better when making an interpretation. Indeed, theories like person-centered are founded on the subjective, the phenomenological; ideas that respond more to working with a client - indeed interpretation is not needed here, as the client has the lead on their own case. If the clinician has perfect data, meaning that the client knows all their information and can plan out their own treatment (to some degree, treatment would not be needed with perfect information).
Continuing with this thought, a client would not enter therapy with perfect information. Perfect information would make therapy useless - it would not be needed because a problem would not exist. Perhaps a problem would exist, though, if the issue (or the solution) is outside the client's possible realm of understanding. In this situation, a therapist is needed. Indeed when incomplete information is had, a therapist is needed. It is possible that perfect information cannot exist if defined as total information - omniscience. No client, or anyone, knows everything. So, does this not mean that 1) a therapist is needed as a side consultant, 2) interpretations (of some kind) are needed, and that 3) pure person-centered therapy can almost never work completely unless interpretation is used in some way.
when they are made. An assumption of a client should never be made on the first session. The first session should be joining and information-gathering. Assumptions come much later in the process, once a clinician has a good idea of the client and his or her issue.