Attribute conditioning (AC) refers to people's changed assessment of stimuli's (conditioned stimuli, CSs) attributes due to pairings with stimuli possessing these attributes (unconditioned stimuli, USs). Up to now, research only showed conditioning of only one attribute within a conditioning session (e.g., athleticism) and measured assessment changes in only this single attribute. The current study shows attribute-specific AC effects in multi-attribute environments and shows that attribute accessibility determines which of a US's multiple attributes are conditioned. Experiment 1 shows AC effects for artificial logos with attributes varying across USs (e.g., athletic, intelligent, or funny). Experiment 2 shows that these AC effects persist over time. Experiment 3 directly manipulated accessibility for attributes varying between and within USs (e.g., a US being sexy and familial) with a priming procedure. Priming-specific attributes prior to conditioning determined which attribute of a US was conditioned to paired CSs. We discuss theoretical implications for AC, as well as practical implications for brand image formation and advertising. Copyright © 2013 John Wiley & Sons, Ltd.