Our unconscious thinking can sometimes be rather sophisticated - with layers of inference and reasoning. This sophistication is particularly evident when the unconscious thiking involves a causal attribution - that is, reasoning about the cause to which we should attribute some fact ("the name rings a bell because it belongs to someone famous").
Consider an early experiment by Nisbett and Schachter. Their participants were asked to endure a seris of electic shocks, with each shock being slightly more severe than the one before. The question of interst was how far into the series the participants would go. Before beginning the series of shocks, some of the participants were given a pill that, they were told would diminish the pain but would also have several side effects: the pill would cause their hands to shake, butterflies in the stomach enz. The pill was a placebo, even so, this inert pill was remarkably effective: Participants given the pill and told about its side effects were willing to accept four times as much amperage as control participants. Nisbett and Schachter proposed that their control participants noticed that their hands were shaking enz (these are standard reactions to electric shocks). The participants then used these self-observations as evidence in judging that they were quite uncomfortable in the experiment. This led them to terminate the shock series relatively early. Placebo participants, attributed the same  physical symptoms to the pill. As a consequence, these participants were less influenced by their own physical symptoms and less influenced by what we called "somatic markers". They detected these markers but discounted them. Of course, participants' reasoning about the pill was entirely unconscious. When they were specifically asked why they had accepted so much shock, they rarely mentioned the pill.