As a parsimonious framework of creative ability, the MTCA is consistent with a number of recurring phenomena in creativity research.

Creativity research is fraught with evidence that intelligence is related to performance on commonly used tests of creative ability, such as divergent thinking, as well as to creative achievements (De Dreu et al. 2012; Jauk et al. 2013; Silvia 2015; Sternberg et al. 2019; see Sternberg and O’Hara 2000 for an in depth discussion of the relationship between intelligence and creativity.). For instance, Silvia (2015) estimates the correlation between IQ and creative ability as measured with divergent thinking tests to be r ≈ 0.50. In addition, tests of intelligence and executive functioning are robust predictors of creative eminence and actual creative achievements (Jauk et al. 2013; Karwowski et al. 2016; Silvia 2015; Simonton 2003a). The idea that cognitive abilities are strongly related is not new. Different subtests of IQ tests show a positive manifold of intercorrelations and different composite scores (factor analytic or sum scores) correlate very strongly (van der Maas et al. 2006). According to this perspective, cognitive abilities that in some creativity theories are considered special for creativity, such as divergent thinking, are simply part of a bigger construct: intelligence1. Although divergent thinking tests are not often part of IQ tests, we see no reason why valid divergent thinking tests with good test–retest reliability (e.g., with more items and automated scoring) shouldn’t be included. In fact, there have been attempts to make this happen (Kaufman et al. 2011; Süb and Beauducel 2005). We note that specific creative achievements may require a different balance of cognitive functions (Murphy 2017). Writing a poem requires more language abilities than solving a chess puzzle. However, because cognitive abilities are strongly correlated (van der Maas et al. 2006), and we use a wide variety of cognitive abilities to analyze, combine, and integrate existing knowledge and skills during creative problem solving, it is hardly surprising that general intelligence tests robustly predict real-world creativity.

Most creativity tests correlate weakly among each other, i.e., have low convergent validity. This could partly be due to the low test–retest reliability (i.e., stability) of standardized creativity tests and resulting error variance (Barbot 2019; Cropley 2000). However, an even more important reason follows from the MTCA: creative achievements build on expertise, and because expertise is domain specific, creative achievements are also domain specific. Whatever the challenge, people simply have to work with the knowledge and skills that they have. An experienced sketch artist with good drawing skills may score higher on the figural subset of the Torrance Test of Creative Thinking (TTCT Torrance 2008) than an experienced creative writer, who, in turn, is more likely to score higher on the verbal subset of the TTCT and the Remote Associates Test (Mednick 1963). The MTCA thus also explains why many experts excel in their profession, but fail to find creative solutions for simple tasks in other domains (Kaufman et al. 2010a). More generally, the expertise component of MTCA clarifies the substandard construct validity of existing creative ability tests (cf. Baer 2012; Han 2003). A prime example of poor construct validity is that the figural and verbal components of the popular TTCT correlate < 0.10 (Baer 2011a, 2011b). Another example is that creativity scores from different domains (e.g., poetry and mathematics) tend not to correlate, or correlate only weakly (Baer 1994a, 1994b; Runco and Albert 1985; Simonton 2003b).

Domain specificity may also explain the low predictive validity of domain general creativity tests, such as the Alternative Uses Task (AUT Guilford 1967). Performance on such creativity tests generally correlates weakly with objective indicators of overt creative behaviors (Kim 2008; Zeng et al. 2011). For example, in a meta-analysis, Kim (2008) found a correlation of 0.22 between divergent thinking test scores and creative achievement. Note that for IQ, correlations with external criteria such as job success are reported to vary between 0.27 and 0.61 (for reviews see Schmidt and Hunter 1998; Sternberg et al. 2001) and predictive values of IQ as high as 0.81 are reported for educational achievement (Deary et al. 2007).

The distribution of achievements is often highly skewed in the population. For example, Murray (2009) describes this for professional golfers. Numerous players have won one or two tournaments, only four have won >30 tournaments, and then only one player (Jack Nicklaus) has won 71 tournaments (Murray 2009). This distribution is referred to as the Lotka curve, and also applies to creative achievements. This curve was first described by Lotka (1926) as a power law function where many authors produce a few publications, but only a few authors produce many publications (e.g., Simonton 2003c). Lotka’s curve has been found to hold for achievements in numerous domains, including chess and music composition (Murray 2009; Simonton 1999c). Many models have been proposed to explain this phenomenon (for overviews see Den Hartigh et al. 2016; Simonton 1999a). Generally, these models invoke some combination of factors, such as latent ability and number of produced outputs. In the MTCA creative ability is a combination of intelligence, which is generally assumed to be normally distributed in the general population, and expertise, which is either normally (in very selective samples) or exponentially distributed. Combining the normal distribution of intelligence and the skewed distribution of expertise results in a skewed distribution, which may explain why the distribution of creative achievements is also skewed (Den Hartigh et al. 2016; Simonton 1999a).

Archival studies show how creative achievements unfold over the career of a creator, be it a scientist, composer, or painter (e.g., Ericsson 1999; Kozbelt 2008; Simonton 1997). What these studies tend to show is that the relationship between the number of creative achievements and career age follows an inverted U-shape, where the number of creative achievements of a creator steeply rises in the early decades of a career, then plateaus and slowly declines. The increase in creativity early in a career can be perfectly explained by expertise development, where the age curve for productivity appears to be a function of career age rather than chronological age (e.g., Khan and Minbashian 2019). Immersion in a domain over time leads to an increase in knowledge, activities, constraints, skills and procedures of a particular domain (Kozbelt 2008). By enriching the knowledge base in a particular domain, the creator has the necessary building blocks for creative work in that domain (Ericsson et al. 1993; Simonton 2008; Weisberg 1999, 2018). However, assuming that creators continue to develop their expertise across their career, why do their creative achievements not show a similar linear increase? Apart from extraneous factors that are beyond the scope of MTCA (e.g., successful scientists get managerial positions that limit their time to express their creative ability), the plateauing of creativity can be explained by how the acquisition of expertise follows an S-shaped curve: after a period of steep increase the acquisition of expertise levels off (Krampe and Charness 2018). That creativity ultimately declines may be caused by various extraneous factors such as poor health or changes in priorities or interest (Kozbelt 2008), but can also be explained by the fact that cognitive abilities, in particular what is generally considered as fluid intelligence, tend to decline during adulthood (Salthouse 2009). Simonton (1997) presented an endogenous explanation of these growth patterns. In future work an elaboration of the MTCA formula in such a dynamic model might be of interest.

Many forms of creativity training exist, most of which focus on the enhancement of a specific ability, such as divergent thinking or convergent thinking (Scott et al. 2004). Many studies show that creativity training effects are generally limited to the abilities that are specifically targeted and hardly generalize to real-world creativity (Baer 2012; J. M. Baer 1988; Scott et al. 2004). This lack of transfer makes perfect sense according to the MTCA. People rely, not on single, but on many cognitive abilities to turn domain specific expertise into real-world creative output. The same studies also show that creativity training effects tend to increase when creativity training is applied to exercises and examples within a relevant domain (Scott et al. 2004). This small, but nevertheless robust strengthening effect, also makes perfect sense according to the MTCA. Real-world creative output builds on domain specific expertise and relevant examples and exercises during training may help people to successfully apply newly acquired skills in their field of expertise.

Decades of research exploring the neural correlates of creativity fail to show a specific brain area involved in creativity (Dietrich and Kanso 2010). If anything, this works shows diffuse prefrontal activation during the performance on creativity tasks (Dietrich and Kanso 2010). In line with this finding, more recent neural models of creativity include a large prefrontal network that is implicated in controlled memory retrieval and central executive processes (Beaty et al. 2016, 2018). This is much in line with the MTCA, which predicts that performance on creative ability tests relies on many cognitive abilities that underlie intelligence in general.