The aim of this paper is to explore the complexity of English noun phrases with emphasis on the order of premodifiers, i.e. to describe the factors that determine the relative position of premodifiers within NP strings. Earlier studies have shown a great deal of variability considering the subject, providing divergent interpretations: structural, semantic, transformational, and psycholinguistic (Feist 2008: 22). Diverse syntactic, semantic, morphological and other properties of premodifying items have impinged upon finding a unitary premodification model by making it difficult to locate a demarcation between their various meanings, interrelations, and usages. Nevertheless, a good deal of systematicity still exists (Langacker 2008: 320). In the paper, we set out to explore this systematicity by conducting a corpus analysis based on the premodification model given by Quirk et al. (1985). The model incorporates multiple linguistic factors in clarifying the phenomenon, which is why we opted for it, hypothesising it would provide a satisfactory explanation of premodifier order. Our analysis features 70 examples of randomly chosen premodification strings obtained by skimming through three different types of sources: an electronic online corpus (Corpus of Contemporary American English, or COCA), Google articles, and four literary works. Once the data had been collected, we proceeded to examine Quirk et al.’s model by exploring how well it fitted our corpus data. The examples that did not fit or may have been described more accurately were processed by additional explication. The results of this study mostly support our expectation—Quirk et al.’s model proved to be effective, successfully incorporating a solid majority of the corpus data. Still, our conclusions should be seen as preliminary pending a more thorough quantitative-qualitative analysis of a bigger corpus of data in the future.