Neological intuition has been defined as the speakers’ metalinguistic ability to identify lexical novelty (Sablayrolles 2003; Lavale-Ortiz 2019; Lombard et al. 2021). In this talk, we present the results of an experiment in which we use lexical decision as an indicator for neological intuition. We address are the following questions and hypotheses: Q1) Is there a difference in recognition time for neologisms w.r.t. attested words / non-words? H1) It is more costly to decide whether a neologism belongs to the lexicon or not; consequently, neologisms are processed slower than both attested words and non-words. Q2) Does the derivational process (suffix) have an effect on the recognition of a neologism? H2) The suffix may have an influence on neological intuition (cf. Lombard et al. to appear). To test these hypotheses, we constructed a lexical decision task using both neological and attested adjectives containing the suffixes -esque, -ien or -ique in French. Although these are globally the most productive denominal adjectival suffixes, they display differences in frequency, availability to express a purely relational vs. qualifying meaning and type of bases selected. The stimuli tested included: -126 neologisms containing one of the three suffixes above. 42 of them were hapaxes (f=1) extracted from the FrTenTen corpus filtered using two common French dictionaries, the rest were unattested forms obtained by crossing the bases and suffixes encountered in the first set (e.g. jungle ‘jungle’ > junglique, hapax; junglesque, junglien non-attested). - the 42 base nouns of these neologisms (e.g. jungle ‘jungle’); - 42 attested adjectives derived with one of the three suffixes above, controlled for length and frequency (e.g. colérique ‘bad-tempered’); - 42 non-words controlled for length (e.g. anudarche); - 42 fillers (e.g. équatorial ‘equatorial’). The experiment was conducted on 67 students of our university (age range 18-27) using the DMDX software. The stimuli were presented one by one on a computer screen. Subjects had to answer the question “Do you think this is a word in French or not?”. Both the answer given (yes / no) and reaction time were recorded. Five base nouns were recognized as existing words in French by less than 75% of the participants (e.g. ballast, geyser). The corresponding derived adjectives were therefore excluded from the analysis. First, as expected, attested adjectives and non-words were recognized or discarded with a good accuracy (93.3% and 94.5% respectively). Neologisms were most often discarded, but with a much lower accuracy rate (74.9%), suggesting a greater deal of incertitude in their treatment. As for the reaction times for “correct” answers (i.e. neologisms identified as non-belonging to the French lexicon), neologisms (=1249,32ms, N=898, =482.57) display a longer processing time than both attested adjectives (=837.4 ms, N=378, =317.56) and non-words (=1162.27 ms, N=378, =496.98). We fitted a linear mixed effect regression to reaction time, predicting it based on length (number of letters) and type of word (neologism, attested or non-word), with participants and items as random effects. Expectedly, neologisms are recognized significantly slower than attested words (Est.:-0.39, p<.0001), but also than non-words to a lesser extent (Est.:-0.08, p<.0001). The shorter the word, the clearer the effect. With regard to the suffix contained in a neologism, it has an effect on accuracy, but not on reaction time. Neologisms containing the more frequent suffix -ique are better identified as attested words (34%) than those containing -esque (22%). However, contrary to our expectations, neologisms containing -ien display the lowest accuracy rate (19%). These observations are statistically significant using a mixed effect logistic regression predicting the participant decision for neologisms as a function of the suffix, with participants and items as random effects. A possible explanation is the fact that our stimuli only contained adjectives based on common nouns, while -ien mostly attaches to proper nouns (cf. Strnadová 2014). Finally, the frequency of the base facilitates the recognition of a derived adjective. This effect is reversed with neologisms, base frequency having an inhibition effect in this case. This could be due to the large difference in frequency between a neologism and its base. Using lexical decision as a measure of neological intuition allowed us to shed light on this notion in at least two respects. First, neological intuition is cognitively costly, since neologisms are slower to process both than attested words and non-words. This confirms the results of previous studies on the subject (e.g. Meunier et al. 2007, Lombard et al. to appear). Moreover, in the case of derived words, the constructional process used may have an influence on the ability of speakers in classifying them as neologisms. References: Lavale-Ortiz R.M. 2019. Bases para la fundamentación teórica de la neología y el neologismo: la memoria, la atención y la categorización. Círculo de lingüística aplicada a la comunicación 80, 201-226. Lombard, A., Huyghe, R., & Gygax, P. 2021. Neological intuition in French : A study of formal novelty and lexical regularity as predictors. Lingua. 254, 103-155. Lombard, A., Huyghe, R., & Gygax, P. to appear. Morphological productivity and neological intuition. Ms. Meunier, F., Dal, G., Fradin, B., & Hathout, N. 200). Rôle de la productivité des procédés morphologiques dans la reconnaissance de lexèmes complexes. In: Morphologie à Toulouse. Actes du colloque international de morphologie 4èmes Décembrettes, München: Lincom Europa: 291-311. Sablayrolles, J. F. 2003. Le sentiment néologique. In L’innovation lexicale. Paris: Honoré Champion, 279‑295. Strnadová, J. (2014). Les réseaux adjectivaux. Sur la grammaire des adjectifs dénominaux en français. Thèse de doctorat. Université Paris Diderot / Univerzita Karlova, Prague.