Native and non-native (even if fluent) speakers recruit different areas of the brain when speaking or comprehending a language. This is now well known from fMRI

Native and non-native (even if fluent) speakers recruit different areas of the brain when speaking or comprehending a language. This is now well known from fMRI studies. http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6T0G-3W1YGT5-9&_user=10&_rdoc=1&_fmt=&_orig=search&_sort=d&view=c&_acct=C000050221&_version=1&_urlVersion=0&_userid=10&md5=4c5ba2d912c6bb17c3920b75ca3a1f51 http://www.sciencedirect.com/science?_ob=ArticleURL&_udi... : From the abstract (Listening to comprehensive but non-native language seems to demand more networked co-processing.) Another one: http://www.sciencedirect.com/science?_ob=ArticleURL&_udi=B6WNP-4J2W0GN-4&_user=10&_rdoc=1&_fmt=&_orig=search&_sort=d&view=c&_acct=C000050221&_version=1&_urlVersion=0&_userid=10&md5=25e0ec3fba6f6d6ce97793cd4ec8353f http://www.sciencedirect.com/science?_ob=ArticleURL&_udi... Non-native speakers recruit more brain regions, suggesting that more "concepts/whatever" are being invoked in the formulation or parsing of a sentence. With this in mind, it would seem that even people who assume they are thinking in a certain language, are only accessing a post-conceptual process (language formulation). Conceptual relations between objects or other concepts may still have been put in place in a ways influenced by the language in which you picked them up. For the same reasons as above, verbal encoding of empathy is probably not the succinct representation the brain uses; you just happen to have conscious access to the verbal encoding only. I would guess though, that something like the Implicit Association Test would uncover more. http://en.wikipedia.org/wiki/Implicit_Association_Test http://en.wikipedia.org/wiki/Implicit_Association_Test As for your Spanish, you will stop using the English-to-Spanish pre-processing crutch once you get more fluent (with practice and repetition).