Have you ever wondered why the English infinitive is sometimes introduced by to and sometimes not? Indeed, have you ever questioned whether English has an infinitive at all? The history of the English infinitive is an interesting one and illustrates an important phenomenon in language development: how speakers of the language recategorize certain grammatical features causing the language to evolve and change.
To answer the second question first, it’s important to distinguish between the infinitive as a form and the infinitive as a function. It’s certainly true to say that modern English does not have a distinct infinitive form. The base form of a verb – the one you find listed in the dictionary (e.g. eat) – can function as an infinitive (I must eat), an imperative (Eat!), the present indicative tense in all but the third person singular (I/you/we/they eat) and as a subjunctive (I insisted he eat). Are these different inflections of the verb that just happen to be identical in form, or a single verb form with a range of functions?
The infinitive of Portuguese verbs has a distinct and easily recognizable form ending in –ar, –er or –ir, the only exception being the verb pôr and its derivatives (dispor, opor etc.), which are only exceptional because the intervocalic n of the Latin verb ponere was lost in the course of time (as also happened in many other words, including the verbs tenere > ter and venire > vir) and the two vowels of *poer were then run together to become pôr. Most of the Romance languages have a distinct infinitive form, as do many other languages across the world. Indeed, in Old English, the infinitive also had a distinct form ending in –an, e.g. etan ‘to eat’, giefan ‘to give’, habban ‘to have’. Two notable exceptions were beon ‘to be’ and dōn ‘to do’.
But Old English also had an inflected infinitive formed by adding the ending –ne to the normal infinitive. This form was used after the preposition to, which in addition to expressing destination also expressed purpose, meaning something like ‘for the purpose of’, so to dōnne meant ‘(in order) to do’. to + infinitive still has this meaning of ‘in order to’ in modern English: to find out more, continue reading.
But why did the Old English infinitive have a special ending after to? The reason is that Old English still had case inflection, just like Latin, German, Russian and many other languages. The preposition to was followed by a noun or pronoun in the dative case which required a special ending. ‘But isn’t the infinitive a verb?’ I hear you asking. In fact, the infinitive is actually halfway between a verb and a noun. Indeed, in Portuguese grammar, the infinitive is referred to as a forma nominal do verbo, ‘a noun form of the verb’, because although it’s basically a verb, it often behaves like a noun. Compare a verbal usage like vou dormir and noun usages like dormir faz bem à saúde, hora de dormir, um estado entre o dormir e o estar acordado. In many languages, the infinitive inflects for case like a noun, and this was also true in Proto-Germanic, the precursor of Old English. The –anne form was the only relic of this inflected infinitive which survived into Old English and it’s clear that, from very early on in the Old English period, speakers understood it to be a verb form rather than a noun. We know this because when it is followed by an object complement in the Old English texts we have, this object is in the accusative case (the usual case for the direct object of verbs) rather than the genitive case (the case used to express the object of a noun, e.g. the prisoner’s release). This shift from noun to verb is an example of grammatical recategorization.
As Old English evolved into Middle English, noun and verb inflections were lost or vastly simplified. Grammatical gender was no longer distinguishable and died out completely, while only a few vestiges of the case system survived, most notably the genitive ‘s ending. The separate dative case forms of nouns had ceased to exist and the infinitive ending had been reduced, first to –en (with the schwa vowel) and then to –e. The combination of these factors meant that the special –anne ending after to also disappeared, especially because once such infinitives were perceived by speakers to be verbs rather than nouns, there was no rhyme or reason for them to have a case ending. The so-called ‘to infinitive’ was born.
It’s not difficult to see why the to + infinitive construction would be the obvious choice for linking a second verb to a first (e.g. want to do, advise someone to do), or to a noun (e.g. chance to do) or adjective (e.g. ready to do), when you consider that the original meaning of the construction in Old English was something like ‘towards doing’, ‘with a view to doing‘, ‘for the purpose of doing’. In fact, the to infinitive can be regarded as the default option in such cases, with only three important exceptions where the so-called ‘bare infinitive’, i.e. the infinitive without to, is required: 1) after modal verbs (can, could, may, might, must, shall, should, will, would) and the auxiliary do; 2) after the causative verbs have, let and make (have/let/make someone/something do) and 3) after perception verbs such as see, watch, hear, listen to, notice, feel, sense etc. (e.g. I saw Bob leave the bar). There are actually a few other cases where the bare infinitive occurs, such as after why and why not (e.g. Why worry about that?), optionally after the verb to help (e.g. I’ll help you (to) learn English), after the expression had better (e.g. We’d better get started) and to avoid repetition of to after and, or, except, but, than, as, like, is and was (e.g. the best thing to do is (to) forget about it).
With the to infinitive being the default option and therefore considerably more frequent than the bare infinitive, to began to be perceived as an integral part of the infinitive. Furthermore, as verb forms continued to simplify into the modern paradigm and the forms of the infinitive, imperative, present indicative and subjunctive became identical, the to started to be categorized as a marker of the infinitive, a kind of replacement for the characterisitic –an ending of Old English, which differentiated the infinitive from other verb forms. As all classes of English words lost their inflections, the to was necessary not only to distinguish the infinitive from finite verb forms, but also to identify it as a verb: there are countless words in English which can be both verbs and nouns (e.g. work, drink, play, to name but three). For this reason, the to is required when the infinitive is used in subject position: (e.g. To ignore the problem would be a mistake). If to was not used in this sentence, the verb would be understood as an imperative and the sentence would therefore be ungrammatical.
Note that this particular to – the one used before an infinitive – had long since stopped being a preposition, being recategorized as a particle, an infinitive marker. We know that it’s different from the preposition to because the latter has to be followed by a gerund, as are all other prepositions in English. Compare I’m used to getting up early (preposition to + gerund) and I used to get up early (particle to + infinitive). Incidentally, English is not the only language with an infinitive marker: a similar phenomenon occurs in the Scandinavian languages (e.g. Danish at spise, Norwegian å spise, Swedish att äta, Icelandic að borða, all meaning ‘to eat’) and even in one Romance language, namely Romanian (e.g. a mânca ‘to eat’).
When the first grammars of English appeared in the 16th century, the description of English grammar was modelled on that of Latin, at that time considered to be the most logical and perfectly structured language. For many hundreds of years, ‘good grammar’ was synonymous with this Latin-based grammar. It was decided that the English to infinitive was the equivalent of the Latin infinitive and therefore not only was to always to be included when quoting the infinitive (that’s why we call it ‘the verb to be’ and not ‘the verb be’), the two elements of the infinitive should not be split. This gave rise to an absurd rule drummed into generations of English-speaking school students: that you should not split an infinitive. It’s an absurd rule because there is no justification for it, but more importantly, when you want to add an adverb that normally precedes the verb to an infinitive, putting it between to and the infinitive is the most natural position and the one we instinctively use in spontaneous speech.
Once grammarians abandoned the notion that to was an integral and indivisible part of the infinitive, it was possible to recategorize English verb forms. Instead of saying that English verbs have an infinitive form, we can say that there is no infinitive form in English and that English verbs have a base form with four functions, one of which is as an infinitive. The role of the particle to is at the same time to connect the infinitive to another element in the sentence and to signal that the verb which follows is functioning as an infinitive. Modal and causative verbs and the auxiliary ‘do’ have to be followed by an infinitive anyway, making to redundant in these cases. Perception verbs can be regarded as similar to causative verbs in this schema, but note that, as in Portuguese, perception verbs can also be followed by a gerund with a slight difference of meaning (cf. I saw Bob leave the bar/Eu vi o Bob sair do bar and I saw Bob leaving the bar/Eu vi o Bob saindo do bar).