I identify the nature and the epistemic status of a sub-type of linguistic intuitions that I call iconic intuitions (IIs). The naive speakers are able to detect, through these intuitions, consistent iconic correspondences between linguistic forms and meanings. Firstly, I identify the main features of the linguistic phenomenon detected by IIs: sound symbolism. The correspondences in which it consists are iconic because they are made up of different types of perceived similarities or associations based on similarities between stimuli - one of which is linguistic. Then, I analyze the main alternative philosophical and psychological characterizations of intuitions, and their evidential role, focusing on linguistic intuitions. On these bases, I conclude that intuitions should be conceived as a heterogeneous construct. Secondly, I argue that the IIs are neither beliefs, dispositions to belief, judgments, or intellectual seemings with propositional contents, but rather perceptual seemings. They consist of the ability or sensitivity to detect iconic correspondences or associations. In other words, sound inputs directly “track” the meanings conveyed by them. They are characterized by their peculiar presentational phenomenology and evaluative component. Now, according to the type of content and cognitive processing involved, it would seem convenient to distinguish between the most purely perceptual ones, based on associative processes, and those that also involve accumulated experience, analytical processes, and conceptual manipulation. After reviewing the psycholinguistic experimental literature based on intuitions about sound symbolism, I argue that IIs are first-level intuitions, and as such a reliable source of direct and prima facie evidence about the iconic features in language. Finally, I argue that these IIs offer a privileged “window” to explore the relationships between language and perception (and affection/emotion). I conclude by arguing that this kind of intuition is a non-dispensable input for philosophical reflection and scientific research on language. So, although I vindicate the relevance of intuitions for understanding linguistic meaning, IIs are not of the same kind, nor do they require the same methods for studying them as those that have mainly interested philosophers.