I have heard it argued, most recently in the entertaining Intelligence Squared Debate “Science Refutes God” (http://www.youtube.com/watch?v=RKNd_S3iXfs), that faith is perfectly rational because we all use it in everyday life, even in science, and so faith in God is no exception. I will argue here that faith in God is a different kind of faith to the most common types of faith that feature in everyday life. In this case we will call faith drawing a conclusion from incomplete data.
First an example from everyday life. The term faith is often used to describe the trust that two people in a committed relationship place in each other. This could mean that they are trusting their partners love and respect for them to continue no matter what. This is definitely a leap of faith, because even in the absence of confirming or dis-confirming evidence of your partners continued dedication then this will be believed. But this not the same as religious faith.
The continuation of trust in their other half regardless of, for example, prolonged physical distance, is a product of years of direct empirical evidence. Loving glances, heartfelt declarations of affection, gift giving, seemingly genuine interest in your life etc. etc. The point is that this kind of faith makes its leap from a reliable and consistent body of data. Religious belief does the opposite. The type of data which religious belief uses as justification for its claims are things like inconsistent historical texts, personal revelation, linguistic arguments and fear. Very little, if no direct empirical observation is used, and the thing which it references (an all powerful omnipotent being) is not a common feature of direct human experience.
On the other hand you have the faith in the partner which is based on evidence gathered directly from the senses about human behavior which is something that almost everyone has much experience of throughout their lives. So that fact that the latter example justifies its faith from a largely dependable source of knowledge (empirical data) and is concerned with something that is a universally common part of human experience (analyzing other peoples behavior) makes it already a more rationally justified faith.
Another difference between the kind of faith featured in the couple example (and countless other common circumstances) and faith in God is that the former will hold trust in the conclusions which the data actually points towards and the latter will hold trust firmly in whatever conclusion they want to believe. A member of the couple makes the kinds of observations of behavior in their loved one as stated above and draws the conclusion that their partner is honest, loves them deeply and so can have trust placed in them. This conclusion is the most likely explanation for the observations made, and so is rational to believe. Even if it turned out their partner turned out to be just a fantastic actor, highly committed to some other gain you offered them, you had no knowledge of this at the time and so because your belief was in line with the available evidence, then it remains rational at that point.
If the data changed nature and began to include unsatisfied glances, decreased will to spend time, increased annoyance/fighting, decrease in declarations of affection etc. etc., then the amount of faith that would be warranted to place in the other half would be reduced. If the amount of faith given stayed constant then then an irrational kind of faith will be held, because the conclusion believed isn’t the most likely given the evidence.
Religious faith is of this kind. Take belief in a personal God that is all knowing, loving and can intervene in human affairs much like the Christian God. Despite knowledge of things like the horrific diseases that have killed billions of innocent people throughout history, insects that lay their eggs in the stomach of other creatures, the many hostilities of the planet we were created, the abhorrent acts that humans have undergone to their fellow kind all through history or the vast, cold, empty chaos of the universe, belief in the kind of God described above prevails. This is a glaring example of when faith is irrational. Even though there is such a massive mountain of features that are inconsistent with Gods infinite loving or her ability to intervene in the universe, the belief in her as such remains wide spread.
This is analogous to, continuing the love examples, someone who has fallen for another person and is pursing them endlessly despite receiving consistent negative responses to their efforts. The receiver of the seduction efforts may be nice to their suitor in some ways, but they are making it clear that they in no way have romantic feelings for them, and may often get very angry and be cruel to them at times when their efforts become irritating and obsessive.
The obsessive lover we would not hesitate to call misguided and irrational, because their beliefs are so out of line with what empirical observations indicate, and I conclude that the same considerations should be applied to judging the rationality of faith in God, even if the small possibility of their being one does turn out to be true. Rationality is about proportioning your beliefs to the available evidence, and so long as God is absent from the evidence, belief in him should be weak at best.