I was listening to an interview with the authors of the new book The Distracted Mind on NPR this morning and they touched on a favorite pet peeve of mine that centers on a basic misunderstanding of the term multitasking. According to Wikipedia, the first published use of the term “multitask” appeared in an IBM paper describing the capabilities of the IBM System/360 in 1965. Is is only recently that the term has been used in the common vernacular to refer to the apparent ability of humans to “concentrate” on more than one task at a time.
One of the authors said something to the effect (and I’m paraphrasing here) that humans cannot multitask because the human brain is not capable of parallel processing. Hold on a second… Setting aside, for a moment, the fact that modern research has shown that information processing in the human brain is immensely parallel, those two terms do not even come close to describing the same thing. The IBM System/360 didn’t support parallel processing, either. There was only one CPU and it was capable of executing no more than one computation at any given time. The term “multitasking” refers to the process of dividing time into small chunks and spreading those small chunks among several tasks, thus apparently making progress on all the tasks at once. This is made possible, of course, by the fact that a computer processes data so quickly that we are unable to perceive the gaps in the processing of any given task and, therefore, observe all the tasks as happening at once.
This is exactly what humans do when they claim they are “multitasking”. It means they are spending small chunks of time on each task such that all tasks appear to be happening in unison while, in actual fact, each task is progressing in fits and starts. So it is incorrect to make a blanket statement about the human brain’s inability to multitask. Of course, anyone who has carried on a conversation with a friend while frying eggs and bacon would know that already.
For some tasks, however, human multitasking doesn’t quite work. The example they gave was trying to listen to the news on CNN while reading the scrolling captions on the bottom at the same time. I admit that if your goal was to parse and understand both threads, that wouldn’t work. But if you were to concentrate on parsing and understanding the scrolling captions while, at the same time, keeping your ears open for any mention of the name “Donald Trump” (or, even better, your own name… should you be unlucky enough to be mentioned on CNN), that should be a fairly simple assignment.
The problem with human multitasking isn’t in the processing power of the brain — the human brain has a huge number of parallel processes running continually from the time one is born until their last breath is drawn (and maybe for a few moments afterward). It’s also not quite a problem of context switching. Sure… humans probably don’t context switch as fast as computers do — but most of the tasks we perform are amazingly slow-moving compared to the speed of the neurons in our brain so while context switching will generally have a noticeable impact on the efficiency with which we manage multiple simultaneous tasks, as opposed to doing them one at a time, most of us are perfectly capable of keeping up with more than one productive (or non-productive) activity at a time.
The real problem is with the size of our short-term memory. The common estimation of human short-term memory is seven concepts, plus-or-minus two. How a “concept” is defined varies. The more data that gets packed into a single “chunk” of information, the more data we can remember short-term. Some tasks require a lot more short-term memory than others. Take, for example, the scrolling captions on CNN. In order to understand what is being conveyed, we have to first store individual words in short-term memory long enough for several of them to be parsed to form a phrase. Then we have to remember the various phrases long enough to combine them into sentences. Then, depending on the content, we may have to remember multiple sentences before we can reduce the content into an understandable concept. The audio feed from CNN is processed in the same way. Seven, plus-or-minus two, is not enough storage to remember all the bits and pieces necessary to simultaneously parse two threads of natural language input. It might be for someone who has practiced doing so for the majority of their lives. In fact, in short bursts (such as during an auction or a heated discussion) we can often manage to parse multiple inputs (though that may have more to do with the ability of our aural processing circuits to record input for a short time and play it back later when the brain has time to parse it). But it’s not something that comes easy to most people.
On the other hand, if most of the multiple tasks in which we’re engaged at any given time have little need for short-term memory, we seem to have no problem multitasking at all. For example, how much do you rely on short-term memory when cooking eggs and bacon? Probably not at all. For most people, that’s an activity that has long since been relegated to the part of our mind called “procedural memory” and we can pretty much do it in our sleep (though, in the literal case of cooking, I wouldn’t recommend it). So the human brain has no problem at all splitting its processing between the mostly-rote activity of frying eggs and bacon and the more short-term memory intensive process of parsing an ongoing conversation with a friend. When only one task requires the use of our relatively scarce short-term memory resources, multitasking isn’t nearly as impossible as popular media would have you believe.
PS: Please do not use this as an excuse for texting while driving. Routine driving can be, for the most part, a rote activity that requires little in the way of conscious processing. But the reason we require a human behind the wheel (at least for the time being) is to manage non-routine events that do require conscious processing and quick reflexes.