Current research on spoken corpora of French reveals that the beginning of sentences may take a variety of distinct forms, which we wish to review in our chapter. Since the concept of “sentence” is unanimously considered as highly unsatisfactory as far as spoken productions are concerned (Berrendonner 1990, Sabio 2006a), the notion of utterance will be preferred here (Pietrandrea et al., 2014 ; Sabio, 2017). Even though most linguists acknowledge the variable nature of utterances in spoken data, they often cling to a fundamental model of “basic sentences” canonically starting with the Subject constituent, possibly preceded by one or more Adverbial phrases (“compléments circonstanciels”, Riegel et al, 1999). Nevertheless, the survey of substantial quantities of spoken productions makes it clear that the “basic sentence” model is seldom observed in actual productions. Our chapter builds on the framework of spoken French description which was initially elaborated in the 1970’s by the Groupe Aixois de Recherches en Syntaxe. The syntactic analysis which we provided is based on the “pronominal approach” framework (Approche pronominale, Blanche-Benveniste et al., 1984), which was subsequently extended to the field of macrosyntax (Blanche-Benveniste et al., 1990). - The pronominal approach is basically a method for describing verbal construction-units which focuses on the matching between syntactic slots (Subjects, Objects, Adverbial phrases) and their pronominal counterparts. The model allows one to distinguish between elements which are governed by a verb (éléments régis) and elements who are linked to the utterance in merely discursive terms (éléments non régis, “ungoverned”). The original pronominal framework was later enriched by a macrosyntactic componant, which aims at describing the utterance-units by distinguishing the central and most informative part (the Nucleus, “le noyau”) and a number of peripheral elements (“satellites”), bearing no illocutionary value. (Berrendonner 1990, Deulofeu 2003, Groupe de Fribourg, 2012). Our chapter wishes to illustrate the fact that the beginning of utterance makes consistent use of some grammatical devices which are diversely exploited by speakers according to the text’s register and “genre”. To that end, we will provide a (non-exhaustive) list of some of the most frequent syntactic features of informal spoken French. Each of the following structures will be introduced throughout the 10 sections of the chapter: 1- The first pages introduce the general framework of sentence / utterance analysis, building on the macrosyntactic approach: the varying degrees of utterance “complexity” will be presented through a set of four “utterance configurations” (configurations d’énoncés). 2- The second section will argue that, although most grammar books usually assume that “in the beginning was the Subject”, Subjects do not always come first in spoken utterances. Some description is provided about subjects’ categories as well as a few distinctive characteristics regarding word-order in unplanned productions. 3- The third section focuses on the peripheral initial elements known as hanging topics, such as la maison (“the house”) in la maison tout est à refaire (“the house everything must be rebuilt”). After describing the most prototypical cases of hanging topics, we consider some more puzzling examples, like those in which the fronted NP is combined with a Subject pronoun (les étudiants ils étaient sympas, “the students they were nice”), and those appearing with verbs like aimer (“to like”) ou connaître (“to know”): Paris je connais pas (“Paris I don’t know”). 4- The fourth section is devoted to the so-called detached Subjects (Sujets disloqués) which are considered as typical of spoken informal language: le sanglier il se nourrit tout seul (“the boars they feed by themselves”). The frequency of the structures is discussed, as well as their grammatical status. 5- The cleft constructions are the topic of our fifth notice: c’est le contributable qui va payer (“it is the taxpayers who will pay”) (Scappini 2006, Rouquier 2018). We present the most frequent types of clefts found in spoken corpora. We then discuss some utterances in which the grammatical status is far less clear, highlighting the specific discursive status of some clefts-like utterances (Sabio 2010, Sabio & Roubaud 2018). 6- The sixth section discusses the forms and uses of pseudo-cleft constructions starting with ce que or ce qui such as (ce qui me plaît c’est la réception de la clientèle, ce que je trouve formidable dans votre livre c’est sa construction, “what pleases me is welcoming customers”, “what I find great about your book is its structure”) (Roubaud 2000, Apothéloz & Roubaud 2015). After a general description of such constructions and an illustration of the most common types, some specific structures are introduced, such as those in il y a, “there is” (ce qu’il y a c’est que le bureau est fermé, lit. “what there is is that the office is closed”) ou faire (“to do”, ce qu’on a fait c’est qu’on a transmis au directeur, “what we did is that we forwarded it to the director”). In our view, pseudo-clefts offer a good example of constructions having one foot in syntactic structure and one foot in discursive patterning. 7- The utterance starting by Si (“if”) and Quand (“when”) are the subject of the 7th section of our chapter (Benzitoun, 2013 ; Sabio, 2013): building on the opposition between the micro- and macro-syntactic level of description, various types are identified, such as “quasi-subjects” or correlative constructions. 8- The 8th section is about Objects which are placed before the verb, in the so-called “Object-fronting” constructions, like dix ans il avait à cette époque (“ten years old he was at that time”). 9- We then address the theme of sentences in which both sequential segments share a specificational – rather that attributive – semantic relationship. Such utterances usually make use of a colon in writing: le problème : il est pas venu (“the problem: he didn’t turn up”), and were consequently termed “colon-effect constructions” by Blanche-Benveniste, 2010. After offering some discussion of such patterns, we provide a tentative typology of the most prominent types in spoken French corpora. 10- Our final chapter questions why some extensively described structures are so seldom used in spoken productions and formulates a few hypotheses on this topic. Key-words: spoken data - syntax - macrosyntax - utterances – French