When you're typing, your senses of touch, hearing, and sight align. You feel, see, and hear your fingers touch the keyboard. Now imagine that you are outdoors and you feel a drop of water hit your hand. If you are like me, then it probably immediately occurs to you that it was a raindrop, so you stretch out your hand to see if more will come, and you look up at the sky for menacing clouds. Let's say the sky is blue and clear as far as you can see. Now your senses of touch and sight are at odds: your sense of touch just told you it was raining, but your sense of sight said it was not. In this case, you don't go running for cover; you choose to go with the information you get from your sense of vision and not the information you got from your sense of touch, probably because you only felt that one drop.
But what if you don't get much more information from one sense than from the other? A team led by Jean-Pierre Bresciani showed people flashes on a screen. At the same time, a device tapped these people's right hand a certain number of times. The screen was set up so that it obscured people's right hand from their field of vision, and the flashes occurred where the right hand would be. People could only feel--but not see--their hand being touched. People were instructed to either count the number of flashes they saw on the screen or the number of taps they felt on their hand, but they were never told to pay attention to both simultaneously. The number of taps given differed from the number of flashes by plus or minus one.
They were never told to pay attention to both flashes and taps, yet apparently, people do automatically pay attention to both. How do we know this? Even though they were told to focus just on the taps or just on the flashes, their accuracy of counting changed whenever the second number of events differed from the one they were focusing on (in other words, the added taps or added flashes messed them up). Take a look at this graph of the results:
Did their counting variability also change? Yes!
One final thing: At the start of the experiment, Bresciani and his team found that when people are only shown flashes or only given taps, their counting is very reliable. When people are shown both flashes and given taps at the same time, there is another interesting thing that happens. Even though you're paying attention to flashes, feeling a tap messes you up. It only sort of works the other way. The effect of touch on vision is more pronounced than the effect of vision on touch. When you're paying attention to taps, vision only rarely messes you up. Seems strange, right? Actually, that's explained by touch being the more reliable sense. We appear to give more weight to the more reliable sense. The fact that vision, the less reliable sense, still affects people's counting of touch, the more reliable sense, means that we automatically process both, but then treat each sense with a weight corresponding to its relative reliability. If we just blocked out the information we get from one sense, then when counting taps, people should not be less accurate when flashes are added, but their count is affected by them, meaning that they do process the flashes.