What are the most common reasons for (synchronic) word-order changes in isolating languages?

Question

What are the most common reasons for (synchronic) word-order changes in isolating languages? From what I’ve read, word order in isolating languages can be changed even when the constituents in the resulting sentences have the same semantic roles as they do in the original sentences.

By “semantic roles,” I mean relationships between agents, patients, etc., with respect to what the verb denotes. I’m trying to follow the SIL’s definition here: http://www-01.sil.org/linguistics/GlossaryOfLinguisticTerms/WhatIsASemanticRole.htm

For example, “The king helped the people,” and “The people were helped by the king” don’t have exactly the same meaning, but the constituents have the same semantic roles. In both sentences, the king is the agent, helping is the action, and people are patients.

I suspect that there are a number of reasons for varying word order in an isolating language.

In Chinese, for example, word order can vary in order to specify whether a given argument is definite. For instance, in Chinese “Lái rén le,” (come person perfect) means “Some person or people have come.” But “Rén lái le,” (person come perfect) means “The people (presumeably the expected people) have come.”

The general rule, if I understand it correctly, is that clause-initial nouns tend to be definite, while non-clause initial nouns tend to be indefinite. See a little more about this Chinese grammar rule here: http://www.chineselanguage.com/chinese-lessons/id/373/grammar--chinese-word-order.aspx

English, my sole and native tongue, is pretty isolating compared to German, but the former has some grammatical constructions that allow for fronting constituents that normally occur after the verb.

To put a focus closer to the front of a sentence, English speakers can do things like this:

We saw Hector in the garden.
a) It was Hector whom we saw in the garden.
b) Seeing Hector in the garden is what we did.

To topicalize non-subjects, we can front them like this:

Rose drove a car.
a) The car, Rose drove.

b) As for the car, Rose drove it.

Unfortunately, the examples I’ve shown here don’t tell me which kinds of word-order changes are the most common across isolating languages. What are the common trends?

Answer 1

It is probably better to talk about the order of constituents rather than words - since it is syntactic constituents that can be moved around, but not single words:

Rose was driving the car.

The car, rose was driving.

*Car, rose was driving the.

You seem to be asking two questions

(1) What is the motivation for word order changes?

(2) What mechanisms of changing word order are used in isolating languages?

I will first try to answer (1):

(1) What is the motivation for word order changes?

Sentences can be analysed, in terms of how they present information, into two parts:

1. What the sentence is about (theme).
2. New information about the theme (rheme)

The king helped the people is a statement about what the king did and might be part of a larger section telling his story: he was intelligent, had to spend his childhood in another country, and after his return, the king helped the people.

But The people were helped by the king is a statement about the people: During the great famine everybody was hungry and desperate. In kingdom A many died, but in kingdom B the people were helped by the king.

Since themes are usually known to those involved in the conversation (discourse-given), they tend to be definite. Usually the subject of a sentence is also its theme because prototypical themes/subjects are high on the animacy hierarchy - such as humans. On top of the animacy hierarchy are those involved in a conversation, who can be referred to with I, we, you, and these are always discoure-given. So it makes sense that subjects, which are often rhemes, usually come at the beginning of a sentence.

(2) What mechanisms of changing word order are used in isolating languages?

Syntactic means such as those referred to by jlawler are available in English and some other languages, such as French. Here's an example of a cleft sentence:

It was Hector I wanted to see, not Juliette.

C'était Hector que je voulais voire, pas Juliette.

In spoken language intonation is at least as important as syntactic mechanisms. Hector receives a pitch accent (major intonation movement) in the following sentences, where Hector is extraposed.

Hector I wanted to see, not Juliette.

Hector je voulais voire, pas Juliette.

Answer 2

Mostly folks try out whatever they think will work in context, and they get to like some ways more than others. I know that doesn't tell you where each one originates, but there's very little direct evidence and therefore most of the possible answers are theory only.

That's partly because there's a lot of syntax in English, and the vast majority of it has been innovated, at least in its present form, during the last millennium, since the grammaticalization wheel turned for English.

For example, here's the table of contents, from p.3 of Haj Ross's list
of The Top 200+ English Transformations of 2012:

I. Emphasis p.3ff
- A. Pseudoclefts and Dislocations pp.3-4
- B. Clefts p.4
- C. Frontings pp.4-6
II. Coordinate Structures p.6
III. Deletions p. 6ff
IV. Noun Modifications p.9
V. Insertions p.10
VI. Pronominalizations p.12
VII. Advancements p.12
- A. To subject p.12
- B. To object p.16
VIII. Ascensions p.18
- A. To subject p.18
- B. To object p.20
IX. Incorporations p.21
X. Rules for Paths p.21
XI. Reorderings p.21
XII. Lexically-governed Deletions p.24
XIII. Embeddings p.27
XIV. Constructions p.28

... and one of the rules, from the bottom of p.5, with example sentences:

C. Frontings
1. TOPICALIZATION:
- [I don’t watch TV ➞ TV I don't watch.]
- [I would never send anything to Harry. ➞ To Harry I would never send anything.]
- [We have been stupid only about citizenship. ➞ Stupid we have been only about citizenship.]
- [We should word this reply tersely. ➞ Tersely we should word this reply.]

NB: Many Southern speakers aver that they can’t do this kind of preposing.
For some reason, NP’s with deictic determiners prepose easilier:
- [[That/This] / > ?The / ?*A] cat I fed.]