Less Wrong is a community blog devoted to refining the art of human rationality. Please visit our About page for more information.
After finding out that old scientists don't actually resist change, I decided to do a literature search to find out if the related assumption was true. Is it mainly just the young scientists who are productive? (This should be very relevant for rationalists, since we and scientists in general have the same goal - to find the truth.)
The answer was a pretty resounding no. Study after study after study found that the most productive scientists were those in middle age, not youth. Productivity is better predicted by career age than chronological age. One study suggested that middle-aged scientists aren't more productive as such, but have access to better resources, and that the age-productivity connection disappears once supervisory position is controlled for. Another argued that it was the need for social networking that led the middle-aged to be the most productive. So age, by itself, doesn't seem to affect scientific productivity much, right?
Well, there is one exception. Dietrich and Srinivasan found that paradigm-busting discoveries come primarily from relatively young scientists. They looked at different Nobel Prize winners and finding out the age when the winners had first had the idea that led them to the discovery. In total, 60% of the discoveries were made by people aged below 35 and around 30% were made by people aged between 35 and 45. The data is strongest for theoretical physics, which shows that 90% of all theoretical contributions occurred before the age of 40 and that no theoretician over the age of 50 had ever had an idea that was deemed worthy of the Nobel prize. Old scientists are certainly capable of expanding and building on an existing paradigm, but they are very unlikely to revolutionize the whole paradigm. Why is this so?
Actually, this wasn't something that Dietrich just happened to randomly stumble on - he was testing a prediction stemming from an earlier hypothesis of his. In "the cognitive neuroscience of creativity", he presents a view of two kinds of systems for creativity: deliberate and spontaneous (actually four - deliberate/cognitive, deliberate/emotional, spontaneous/cognitive and spontaneous/emotional, but the cognitive-emotional difference doesn't seem relevant for our purposes). Summarizing the differences relevant to the aging/creativity question:
According to this framework, insights can occur during two modes of processing, deliberate and spontaneous. Deliberate searches for insights are instigated by circuits in the prefrontal cortex and thus tend to be structured, rational, and conforming to internalized values and belief systems. Spontaneous insights occur when the frontal attentional system does not actively select the content of consciousness, allowing unconscious thoughts that are more random, unfiltered, and bizarre to be represented in working memory. Several lines of evidence corroborate the notion that deliberate insights are different from spontaneous insights. For instance, the prefrontal cortex is recruited in long-term memory retrieval (for reviews, see Cabeza & Nyberg, 2000; Hasegawa, Hayashi, & Miyashita, 1999) and thus can be said to have a search engine that can “pull” taskrelevant information from long-term storage in the posterior cortices, momentarily representing it in the working memory buffer. Once online, the prefrontal cortex can use its capacity for cognitive flexibility to superimpose the retrieved information to form new ideas. (...)
...suggesting that solutions that would violate what is known about the world are not readily considered in deliberate creativity. Moreover, the prefrontal cortex houses a person’s cultural values and belief system (Damasio, 1994). Thus, the processes of effortful retrieval and recombination of knowledge yield results that are highly consistent with a person’s world view and past experiences (see Dietrich, 2004). Another critical limitation of the deliberate processing mode is due to the fact that any information that is retrieved deliberately and is thus explicitly available for conscious manipulation is subject to the capacity limit of working memory. (...)
In contrast, the spontaneous processing mode produces insights that are different qualitatively because they are not initiated by prefrontal database searches that are limited to preconceived mental paradigms as well as quantitatively because information is not subject to the capacity limit of working memory. During the inevitable times when the frontal attentional system is downregulated, for instance in daydreaming, thoughts that are unguided by societal norms and unfiltered by conventional rationality become represented in working memory (Dietrich, 2003). In such a mental state, conscious thinking is characterized by unsystematic drifting, and the sequence of thoughts manifesting itself in consciousness is more chaotic, permitting more loosely connected associations to emerge. (...)
The prefrontal cortex is the last structure to develop phylogenically and ontogenically (Fuster, 2000, 2002). In humans, it is not fully matured until the early 20’s, which is likely the reason why the creativity of children is less structured and appropriate. Likewise, evidence suggests that prefrontal functions are among the first to deteriorate with age. Data from humans and other animals show that aging individuals are less able to inhibit well-learned rules and have less independence from immediate environmental cues or long-term memories (e.g., Axelrod, Jiron, & Henry, 1993; Means & Holstein, 1992). This tendency to adhere to outdated rules might be compounded by the fact that mental states that enable the spontaneous processing mode, such as REM sleep or daydreaming, go dramatically down with age (Hobson et al., 2000; Singer, 1975). Thus, in addition to perseveration, the deliberate processing mode, which favors solutions that tend to be consistent with a person’s belief system, becomes the more dominant problem solving mode of thought in the elderly.
So, it seems like the older we get, the more likely it is that our thinking is dominated by pre-conceived ideas. This isn't automatically a bad thing, of course - those "pre-conceived ideas" are the ones we've been building for our whole lives. But it isn't good if that prevents us from coming up with entirely new yet good ideas. The empirical evidence seems to suggest it does.
What can we do to combat this? Different cognition-affecting drugs are one answer that automatically springs to mind, but many of those are for a large part both illegal and unsafe. Maybe we should try to spend more time daydreaming the older we get, or explicitly using our cognitive creativity to try to generate ideas which smack to us as senseless at first? But there are far more ideas that both seem and are senseless to us, than there are ideas which seem senseless and actually aren't, so the low hit ratio may be pretty exhausting.