Preposition stranding

• 1. Introduction

  ›

  Preposition stranding is the grammatical phenomenon that has sparked centuries of argument among writers, teachers, and linguists about what English speakers are and are not allowed to say. The question at the center of that argument is deceptively simple: can a preposition be left hanging at the end of a sentence, separated from the object it belongs with? That is the sort of question grammarians have wrestled with. And that last sentence is itself an example of exactly what is being debated.

  The term "preposition stranding" was coined in 1964, though the slightly different phrase "stranded preposition" had already appeared in 1949. Before those terms existed, linguists described the same construction as a sentence-terminal preposition. The phenomenon turns up in English and other Germanic languages, as well as in two Niger-Congo languages, Vata and Gbadi, and in certain dialects of French spoken in North America.

  What makes preposition stranding so interesting is not just its grammar but its history. Since at least the 17th century, writers and schoolteachers have tried to outlaw it, while the people those teachers and writers were trying to correct have gone right on using it. This story involves a poet, a grammarian, a historian, and a quip famously attributed to Winston Churchill.

• 2. When Grammarians Objected

  ›

  The earliest recorded criticism of stranded prepositions can be traced to the 17th-century grammarian Joshua Poole. The controversy gained real momentum in 1672, when the poet John Dryden objected to a phrase that the playwright Ben Jonson had written back in 1611: "the bodies that those souls were frighted from." Dryden thought the preposition should come earlier in the sentence rather than sitting at the end. What he never did was explain precisely why.

  Dryden's own earlier writing had used terminal prepositions freely. He went back through later editions of his work and removed them, and he explained his reasoning in an unusual way: when unsure about English usage, he would translate his sentences into Latin to test their elegance. The logic was that Latin lacks any construction comparable to preposition stranding, so if the Latin version sounded wrong, the English version must be wrong too.

  Robert Lowth picked up the thread in his 1762 textbook, A Short Introduction to English Grammar. His verdict was careful rather than absolute. He wrote that stranding was "more suitable for informal than for formal English," calling it "an Idiom which our language is strongly inclined to" and suitable for "common conversation" and "the familiar style in writing." He acknowledged that fronting the preposition was "more graceful" and better suited to solemn writing. Lowth himself used the construction in that very passage, including in the phrase "is strongly inclined to," which sits at the end of the sentence he used to describe the rule.

  By the 19th century, the proscription had found its way into English school teaching, and it was still being taught in some schools at the start of the 21st century.

• 3. Defenders and Rebuttals

  ›

  H. W. Fowler pushed back hard. In his Dictionary of Modern English Usage, published in 1926, he called the ban on terminal prepositions a "cherished superstition" that ran against what he described as the "incurable English instinct for putting them late." He devoted four full columns of his Dictionary to dismantling the prescription.

  Fowler's argument was that the rule, when applied too aggressively, produced uglier sentences than the thing it was trying to prevent. He illustrated this with the example "That depends on what they are cut with," and showed how insisting on formal preposition-fronting could turn it into the grotesque construction "That depends on the answer to the question as to with what they are cut."

  The historian Edward Gibbon took the opposite extreme. He not only avoided sentence-terminal prepositions but extended his avoidance even to phrasal verbs, steering clear of putting "on," "over," or "under" at the end of a sentence, even when those words were clearly functioning as adverbs rather than prepositions.

  The American linguist Don Ringe identified the original objection as a category error. He pointed out that critics were insisting English grammar should resemble Latin grammar, and he called that logic "perverse," noting that English is not Latin and is not descended from Latin. The quip attributed to Winston Churchill put the same point in sharper and funnier terms: "This is the sort of tedious nonsense up with which I will not put." The sentence is technically correct under the strict rule, and it sounds absurd.

  As O'Conner and Kellerman note, great literature from Chaucer to Milton to Shakespeare to the King James Bible was full of what critics called terminal prepositions. Mignon Fogarty, writing as "Grammar Girl," sums up where most sources have now landed: "nearly all grammarians agree that it's fine to end sentences with prepositions, at least in some cases."

• 4. How Stranding Actually Works

  ›

  Preposition stranding does not happen in just one way. It arises in at least three distinct grammatical situations: wh-movement, sluicing, and what are called pseudopassives.

  Wh-movement involves question words like who, what, when, where, why, and how. In English, asking "Which town did you come from?" leaves the preposition "from" stranded at the end. The alternative, "From which town did you come?", carries the preposition forward with its object, a process linguists call pied piping. English and Dutch permit both options. In Serbo-Croatian and Arabic, pied piping is the required choice, and stranding is not allowed. Greek also requires pied piping in wh-movement. In Danish, stranding is banned only when the wh-word is in the nominative case; accusative cases such as the equivalent of "whom" permit it.

  Sluicing is a different animal. It is a form of ellipsis in which a wh-phrase stays in place while the rest of a clause is deleted. English allows preposition stranding under sluicing; the example from the source is "John laughed at someone, but I don't know who he laughed at." The theory known as the preposition stranding generalization, or PSG, proposes that any language allowing stranding under wh-movement will also allow it under sluicing. That rule does not hold universally.

  Pseudopassives, also called prepositional passives, arise when the object of a preposition moves to fill the subject slot in a passive construction. The example given in the source is "This bed looks as if it has been slept in." In this case, the stranded preposition is not the result of a wh-word moving but of a noun phrase shifting to subject position.

• 5. Across Languages

  ›

  Spanish allows no room for prepositional stranding. Pied piping is the only grammatical option when forming oblique relative clauses in Spanish, which means the preposition must travel with its object rather than be left behind.

  French is more complicated, because the answer depends on which French you are speaking. Standard French requires pied piping in most constructions. A speaker of standard French encountering dialect examples of stranding would, according to the source, regard them as barbarisms or anglicismes. Yet Prince Edward Island French permits stranding in certain wh-question contexts. Ontario French restricts stranding to relative clauses and only with certain prepositions; stranding is impossible in most dialects when the prepositions involved are "a" (meaning "to") or "de" (meaning "of").

  Arabic splits along regional lines. In Emirati Arabic, stranding in wh-movement is possible but only through a specific construction using which-NPs followed by IP-deletion. Libyan Arabic normally bans stranding in wh-movement sentences, though a recent study found that a preposition appears to be stranded in resumptive wh-questions in that dialect.

  Dutch offers yet another variant through its r-pronouns. Dutch prepositions do not generally take ordinary neuter pronouns like "het" or "dat" as objects. Instead, they combine with corresponding r-pronouns such as "er," "daar," and "waar" to form single words, giving "erover" rather than "over het." When r-pronouns are moved to the left in a sentence, the postposition they were fused with gets left behind. German dialects show a parallel pattern with da(r)- and wo(r)- forms, a construction those dialects call a Spaltkonstruktion, or split construction. Standard German requires the composite form; the dialects allow the split.