- PREPUBLICATION VERSION -

Preposition Placement in English. By Thomas Hoffmann. Cambridge: Cambridge University

Press, 2011. Pp. 297. Reviewed by Hans C. Boas, The University of Texas at Austin.

Which factors influence variable preposition placement in pied-piping (“About what is he

talking?”) and preposition stranding (“What is he talking about?”)? Are the factors the same
across different clause types? Are there any differences in different varieties of English, such as
British English and Kenyan English? Which formalism is best suited to capture the extensive
data on preposition placement in English? Drawing on corpus-based as well as experimental
data, Preposition Placement in English offers a wealth of insightful answers to these questions.

Chapter 1, ‘Introduction’, provides a brief overview of the phenomenon and introduces some of
the key ideas that have led to writing this intriguing book, including the need for corroborating
evidence, usage-based approaches, processing considerations, the status of different varieties of
English, and the importance of arriving at a solid description of the data before arriving at a
formalization. One of the many strengths of Hoffmann’s book is that it allows researchers with
different theoretical backgrounds to access his detailed analyses of the data. The chapter sets the
stage for his work in that it makes explicit that the results of his data analysis are presented in
theory-neutral terms first - only his final chapter offers a particular formalization employing an
HPSG-like constructional approach (Head-driven Phrase Structure Grammar; Pollard and Sag
1994, Ginzburg and Sag 2000). Chapter 1 ends with an overview of the book.

Ch. 2, ‘Corroborating evidence: Data and methodology’, offers one of the most refreshing
discussions about the role of evidence in linguistics that I have read in recent years. Hoffmann
does a fantastic job of juxtaposing two rather opposite views in modern linguistics with respect
to the role of employing corpus data versus introspection data. Discussing advantages and
disadvantages of both approaches, Hoffmann addresses all important points, such as the role of
representative corpora, negative evidence, intra- and inter-subject consistency, and objectivity.
His discussion leads him to argue for the use of multiple sources of different types of data, which
he terms ‘corroborating evidence’ (similar to Gries, Hampe and Schönefeld’s 2005 concept of
converging evidence). In staying with the metaphor from the semantic domain of crime,
Hoffmann explains in great detail the role of corpus data (“Exhibit a”) and introspection data
(“Exhibit b”) for his work. With respect to the former, Hoffmann first presents an introduction to
the International Corpus of English (ICE) corpus (design, tagging, extraction of data), followed
by an in-depth discussion of a variety of modern “forensic tools,” including programs used for
statistical analysis such as Goldvarb, R, Coll.analysis 3, and HCFA. With respect to the latter,
Hoffmann offers an expertly overview of experiment design, magnitude estimation experiments,
and statistical programs to analyze introspection data. I found this chapter to be extremely useful

1
as it gives readers a superb overview of different methodologies as well as tools to collect and
analyze data from various perspectives. This chapter would make an excellent read for a
beginning graduate course on analyzing linguistic data.

In Ch. 3, ‘Case notes: Independent factors’, Hoffmann unravels the complex set of factors to be
analyzed and weighed in his study of preposition placement in British English and Kenyan
English. The chapter begins with construction-specific constraints on preposition placement in
structures such as interrogative, exclamative, wh-relative, non-wh-relative, comparative, hollow,
and passive. Each clause type receives a detailed discussion as to the constraints on preposition
placement and the interaction effects, including an expertly review of the previous literature and
a brief comparison with data in German which is much less flexible with respect to preposition
placement. I most enjoyed the heart of Chapter 3, which deals with different variable effects
across clause types, in order to answer the question of whether individual clause types also
exhibit idiosyncratic variable constraints on preposition placement. Hoffmann’s illuminating
discussion of this matter is rounded off by a discussion of the diachronic and typological
evidence. The remainder of this very long chapter (77 pages total) focuses on other important
issues, including types of PPs, level of formality, PPs embedded in NPs and AdjPs, complexity,
and second language effects (Kenyan English). In addressing these issues, Hoffmann does not
only an admirable job of presenting the latest thinking of the topic, but he uses his discussion to
foreshadow important points that are of relevance for his own analysis in subsequent chapters. At
the end of Chapter 3 I felt like I had a very good understanding of the complexity of preposition
placement in different varieties in English. The only drawback of chapter 3, in my view, is the
relatively short discussion of L1 transfer in Kenyan English. While a full discussion of all the
relevant factors involved is probably out of the scope of Hoffmann’s study it would have been
nice to provide more references on L1 transfer in general (see, e.g. Clyne 2003), and to offer the
reader a more detailed outline about the criteria one could employ to decided whether L1 transfer
did in fact take place or not.

Ch. 4, ‘Evidence I: Corpus results’, presents the results of Hoffmann’s statistical analyses of the
data from the ICE-GB (Great Britain) and ICE-EA (East Africa) corpora. I found this chapter an
excellent read, even though some of the discussion and interpretation of the statistical results
were difficult to follow for someone not intimately familiar with the details of statistical analysis.
One of the most influential insights coming out of Hoffmann’s work is perhaps that preposition
placement in Kenyan and British English is influenced by several similar factors while at the
same time exhibiting some idiosyncratic effects – each illustrated by a wealth of data, neatly
organized in tables throughout the chapter. The chapter begins with a discussion of the coding of
the ICE data with six contextual factors, together with an explanation of how the Goldvarb
program helps with analyzing the data. The remainder of the chapter discusses in detail
categorical clause contexts, variable clause contexts (ICE-GB vs. ICE-EA), and relative clauses.
The wealth of the data and comprehensive insights cannot all be summarized in such a short

2
review. Instead, I would like to point out that Hoffmann not only does a superb job with
analyzing and presenting the data. In addition, he also offers a lucid debate about how the data
can be interpreted in different ways, thereby presenting the reader a variety of different options
without settling on one option at the beginning. For example, on p. 129 he begins with
investigating the factors influencing the distribution of stranded prepositions in categorical
clausal contexts. Subjecting British and Kenyan data to an HCFA analysis, he determines the
status of the variables VARIETY (British vs. Kenyan English), PP, and CLAUSE (finite non-wh-
relative clause, free relative clause, passive, hollow, and comparison clause). The results of the
HCFA analysis are presented in a variety of tables illustrating the configurations of the
individual variables, the configurations of all two-way interactions, and the configurations of all
significant three-way interactions. In discussing the data, Hoffmann is careful to not come to any
premature conclusions, but offers the reader various options of how the data could be interpreted.
At each step, Hoffmann then presents more data to show how the different factors interact with
each other. The beauty of this corpus-based chapter is that the data speak for themselves, while
Hoffmann only appears to be guiding the reader through the maze of information.

Ch. 5, ‘Evidence II: Experimental results’, presents the results of Hoffmann’s introspection
experiments using the Magnitude Estimation method. Based on the experimental data collected
among speakers of British and Kenyan English, Hoffmann first demonstrates that pied-piping
and stranding are both more entrenched in British English simple relative clauses. In contrast, the
Kenyan English data show that the second-language variety favors the prototypical association of
stranding with prepositional verbs and pied-piping with adjunct PPs, which requires less
processing efforts, according to Hoffmann. These data corroborate his corpus-based findings and
should be taken as another advantage of using different types of data and analyses. Subsequent
sections in Chapter 5 address preposition placement in relative clauses of various complexity and
preposition placement in interrogative clauses. Similar to Chapter 4, Chapter 5 presents a wealth
of data and tables, together with an easy-to-read guide as to how to interpret the data. A complete
summary of Chapter 5 would fill a whole page so I will focus on the (in my view) two most
important ones: (1) pied-piping with that- and 0-relativizers is a hard grammatical constraint in
British and Kenyan English; (2) With increasing complexity (of relative clauses), the stranding
preference for prepositional verbs is increased in both varieties (p. 224).

Ch. 6, ‘Preposition placement: The case for a Construction Grammar account’, offers a precise
formalization of the insights of previous chapters. The chapter explains Hoffmann’s choice for
Construction Grammar over other linguistic theories: it offers both descriptive and explanatory
adequacy (see Goldberg 2006). Since Construction Grammar is not a coherent linguistic
framework, but comes in many different flavors which all see the concept of “construction” (a
pairing of form with meaning) as central to linguistic analysis (see Trousdale and Hoffmann
2012), Hoffmann is in an enviable position to essentially combine the advantages of two
Construction Grammar variants into one, namely HPSG Construction Grammar (Ginzburg and

3
Sag 2000), together with usage-based Construction Grammar (Goldberg 2006). While the
former, which in the meantime has been developed into Sign-based Construction Grammar (see
Boas and Sag 2012), is focused on providing maximally precise formalizations of linguistic
insights, the latter is also interested in offering psychologically plausible explanations about the
motivation of linguistic structures. Hoffmann’s short introduction to Construction Grammar is
followed by a sketch of a minimum complete-inheritance network seeking to capture preposition
placement. Based on a short introduction of the main principles of HPSG and a discussion of
Sag’s (1997) analysis of English relative clause constructions, Hoffmann formalizes and
discusses all relevant constructions that are a part of the complete-inheritance HPSG-based
Construction Grammar account of preposition placement in British and Kenyan English. The
level of Hoffmann’s attention to detail is astounding, as is his ability to present fairly complex
material in a straightforward fashion. At the end of this chapter, I was left with the desire to see
more studies like Hoffmann’s, which combines detailed theory-neutral analysis of a variety of
different data with a level of explicit formalization that is unfortunately rather rare these days.

All in all, the book is an excellent read. There are only a few typos (“GB” should be eliminated
at the top of Table 4.23, p. 161; “the” is missing in the third line of the third paragraph on p. 226;
“possess” on p. 275 has an “s” missing). Obviously, no study is ever complete, and Hoffmann’s
book is no exception. While he offers a wealth of data, a detailed analysis, and a precise
formalization of his observations, I was sometimes left with the impression that there could be a
bit more of an explanation of the facts that go beyond a description, statistical analysis, and
formalization. For example, on p. 132 Hoffmann discusses the results of the three-way
interaction configurations, leading him to claim that they “can again be considered to reflect the
cognitive entrenchment of particular constructions.” He then goes on to discuss the differences
between British and Kenyan English, but at the end of his discussion I was left wondering as to
why this entrenchment had taken place? In other words: Which linguistic (or more broadly
speaking, cognitive) factors lead to this cognitive entrenchment? How do we know? What does
that tell us about not only the nature of this construction, but also about the linguistic system as a
whole? In my reading of Hoffmann’s book such questions came up repeatedly, especially in
chapters 4 and 5. This issue should not be regarded negatively as it is simply not possible to offer
an even more complete account of a phenomenon as complex as preposition placement in
English. All in all then, Hoffmann’s book is a pleasure to read and the arguments are easy to
follow. I recommend it highly to anyone who has ever wondered about preposition placement in
English.

References

Boas, Hans and Ivan Sag (ed.). 2012. Sign-based Construction Grammar. Stanford: CSLI
Publications.
4
Clyne, Michael. 2003. Dynamics of language contact. Cambridge: CUP.

Ginzburg, Jonathan and Ivan Sag. 2000. Interrogative Investigations: The Form, Meaning, and
Use of English Interrogatives. Stanford: CSLI Publications.

Goldberg, Adele. 2006. Constructions at work. Cambridge: CUP.

Gries, Stefan, Beate Hampe and Doris Schönefeld. 2005. Converging evidence: Bringing
together experimental and corpus data on the association of verbs and constructions. Cognitive
Linguistics 16, 635-76.

Pollard, Carl and Ivan Sag. 1994. Head-driven Phrase Structure Grammar. Chicago: University
of Chicago Press.

Sag, Ivan. 1997. English relative clause constructions. Journal of Linguistics 33, 431-484.

Trousdale, Graeme and Thomas Hoffmann (ed.). 2012. The Oxford Handbook of Construction
Grammar. Oxford: OUP.