The sample from the link below provides some evidence against this occurrence:

"If the bond between the verb and particle is too tight, following transformations are ruled out for phrasal verbs, the particle cannot undergo fronting(14)...

(14) a. Out jumped the frog! (Cappelle 2004:39)

b. *Up he made a story. (Darwin, Gray 1999:78)"

I believe this isn't possible in English, and as a native speaker, my instinct wants to confirm this.

http://www.grin.com/en/e-book/140783/between-heads-and-phrases

Article Going After the Phrasal Verb: An Alternative Approach to Cla...

Thanks! I worked on German particle verbs and they are very heterogeneous. Basically everybody claimed that particles cannot be fronted but there is attested data that shows that there is a broad variety of particle verbs that do allow fronting. So maybe we find some English example as well. =:-)

You may want to consider cleft structures; more examples can be found in prose and informal chats.

The BNC seems to contain plenty of examples, ie searching for sentence initial particle followed by a finite verb or personal pronoun (to filter false hits)

off he goes

up goes the axe

etc

Hi Stefan,

I found this in a lyrical context (sixth verse on the following page): It's a page for kids to get smart. And what's good enough for kids can't be bad (or ungrammatical) for adults.

Hi Stefan, 

I wrote something about this in Cappelle (2002), "And up it rises: Particle preposing in English" and in my PhD thesis too (2005: "Particle Patterns in English"). Preposing with a transitive verb is generally rather bad, but is perhaps marginally acceptable if the subject is an unstressed pronoun. I'm quoting from my own thesis here:

(p. 230) Particle preposing has been claimed in the literature to be restricted to intransitive verb-particle combinations (Palmer 1965: 184-86; Jackendoff 1973: 346; Quayle 1994: 254). This is not entirely correct. It is true that the great majority of all particle preposing instances involve an intransitive combination, but transitive verb-particle combinations are allowed in this pattern as well. Here are three examples (italics added):

(366) a. They saw the survivors required dry and warm clothing, so off they took them to their own cabins to fit them out with everything they could. (www example; URL lost, accessed around 4 June 2004)

b. About ten o’clock there was a flag of truce and in they brought the Captain and three men. (my.execpc.com/~cra1/Roblet1.html, accessed 4 June 2004)

c. Wood and I boarded the Vixen, and there we got Lieutenant Sharp’s black Cuban pilot, who told us he could take our transport right in to within a few hundred yards of the land. Accordingly, we put him aboard; and in he brought her, gaining at least a mile and a half by the manoeuvre. (www.sonshi.com/theodore_roosevelt2.html, accessed 4 June 2004)

Occasionally, the verb-particle is only intransitive ‘on the surface’, being in the passive voice:

(367) At that moment in was brought a loin of stag and other venison in great plenty, and rich golden plate adorned the table, with great lidded goblets of gold, and magnificent golden candlesticks bearing great candles 

(www.uidaho.edu/student_orgs/arthurian_legend/grail/fisher/texts/romance/perlesva.htm, accessed 4 June 2004)

Preposing is only possible if the subject refers back to a familiar entity and is light. Compare the previous sentence with the following sentence (discussed in Cappelle 2002: 64):

(368) *At that moment in four waiters brought (or in brought four waiters) a loin of stag and other venison in great plenty…

The explanation for the unacceptability of this sentence probably has to be sought in a restriction against the consecutive placement of informationally rich elements. Compare with (366)b, where the two focused elements (the  (p. 231) particle and the subject NP) are widely separated in the clause: one at the front and the other at the end. 

See you!

Best,

Bert

Well, Rui, you could call them "adjectives of sorts" if you like, but they're certainly not true adjectives. Okay, on occasion they might be used after the verb be, but that's about it. You can't use them attributively (*the off clothes, *the away book, *the back dog, etc.) and even coordinating them after be with uncontroversial adjectives would be weird (*?the gloves are superfluous and off). 

But you're absolutely right that if the particle forms an idiomatic combination with the verb, it doesn't have an independent meaning, so fronting it wouldn't make sense then. 

Hi Bert,

Thanks for your answer! The reason for asking the question was your paper about allostructions, which I just read. If you say that verb object particle combinations are phrasal constructions, then there seems to be a problem with the extraction of the particle, right? The extracted particle should be related to the verb. Usually this is done by some sort of unbounded dependency. You seem to be forced to assume another allostruction for fronted cases or traces in the allostructions you already have in your paper. Both solutions seem problematic to me.

Does your thesis contain a treatment of these frontings that is compatible with your later allostructions paper?

Thanks again!

Greetings from Berlin

Stefan

Hi Gosse,

Thanks! But your screenshot only shows intransitive cases. I will also have a look.

Greetings

Stefan

Hi Stefan,

I do indeed believe that V-NP-Prt is a phrasal (a VP) construction, while in V-Prt-NP the V and the Prt form a word-level unit. Both configurations are, as I see it, possible grammatical manifestations of a single lexical unit which itself is underspecified for its precise grammatical status (syntactic or morphological). In this allostructional model, the Prt is neither extracted from the V (which would assume V-Prt-NP) nor incorporated into it (which would assume V-NP-Prt) is basic. In both manifestations, the Prt is related to the V, independently of its exact position. (There is a tendency, though, for idiomatic particles to prefer the position adjacent to the verb, but as you know, we can also perfectly say I forgot to back my files up, for instance).

If one accepts this admittedly rather simple solution (no derivation, just two independently existing structures that fall under a more general [V + Prt]trans schema, then one could also accept the existence in grammar of preposing structures. In my thesis, I have the following structures:

a.  [S PrtP V NPSubj]

b.  [S PrtP NPSubj; light and discourse-old V (Comp to Prt)]

Note that in these structures, the particle can project a full phrase. Its optional prepositional complement can be in the canonical post-verbal position though, as in:

The rain stopped so back out we went to the main stage and caught the end of Supersuckers… (www.jimtingle.eclipse.co.uk/leeds2K.html, accessed 7 April 2004)

So here, I treat out as the head of the particle phrase, with back as some sort of specifier and to the main stage as a complement. The details do not matter, but it may be interesting to see that as the particle is preposed, it obviously cannot form a morphological unit with the verb in that structure; it has to be a clausal constituent and in this capacity it can take right for instance, just as in the V NP PrtP allostruction.

For transitive V+Prt combinations, then, I suggested the following structures, which fit (366b) and (366a, c) in my previous answer:

a.  [S PrtP NPSubj; light and discourse-old V NPObj]formal

b.  [S PrtP NPSubj; light and discourse-old V NPObj; light and discourse-old (Comp to Prt)]formal

There's nothing very deep about this analysis, which you might argue just lists the structures without really explaining how they are licensed by independent grammar rules. All I can say, in retrospect, is that the proposal still strikes me as intuitively right, at least to the extent that if the particle can form a complement of the verb (while still being semantically related to it!), it can form a phrase and it can be preposed, just like any other complement. Since the preposed position is for focal or topical constituents, it's only natural that only particles that by themselves can "hold" such a focal or topical position can be found in these additional constructions; this is why idiomatic particles are excluded.

So, in short, my proposal is to just proliferate allostructions!

Greetings from... Germany, as it happens,

Bert

Stefan, 

I missed the 'transitive' part of the question.

Indeed, I have not been able to find cases with a transitive verb in either BNC or COCA. (But note again, that only approximate queries are possible.)

Hi Rui,

Oh yes, I fully agree with this point. I like to think of parts of speech as pretty stable categories, but they are of course abstractions. There are lots and lots of lexical items that don't find themselves comfortably right in the middle of any of these categories but rather straddle the boundaries between them (if you can talk about boundaries at all). Another example would be "many", which can be thought of as a determiner ("many books") or as an adjective ("the many books"). I just thought you wanted to rename spatially used particles as "adjectives", but I think you rightly added "of sorts". So yes, I certainly see no harm in saying they can take on "adjectival properties". This probably means that we should think of parts of speech as bundles of semantic and distributional properties. 

I would look into contrastive focus constructions... maybe something like:

A: Did you say he threw the tuna out?

B: Up he threw it, not out!

Not that this pair is a particularly felicitous example, but maybe it points in a useful direction.

Some seem better to me than others.  I haven't given it much thought about which classes of particles can and which can't be fronted (perhaps a distinction between particles acting as adverbials and those that encode resultativity?).  The fronted ones work better with emphasis on the fronted particle.  

Intranstiives seem to do this fairly productively.

(a) Off he walked with the money vs he walked off with the money (not transitive, I know)

(b) Away he flew into the sunset vs he flew away into the sunset

(c) I told him to come along and sure enough, Along he came

(d)  I thought I was the first to tee off but he beat me to it and  off he tee-ed

(e) Out they wanted to go and Out they went 

But not:

(f) *Up he threw vs he threw up  (but this may be a covert transitive because one can insert a pronoun "it" here)

(g) *I told him to keep fighting but in he gave vs he gave in

Transitives

(a) Away he kicked the ball vs he kicked the ball away

(b) *Out, he threw the trash vs he threw the trash out

(c)  *off he pissed me vs he pissed me off

(d) *Up he brought it vs he brought it up

(e) *down he played it vs he played it down

(f) *Out he took her vs he took her out

There is a potentially useful contrast between transitive (f) and intransitive (e).  (e) above is grammatical but only on the literal reading of "going out" to an event or a party.  It is ungrammatical on the reading that they were dating.  So this is a nice minimal pair.    This seems to be supported by  transitive (f) which, in the unfronted form, only has a reading where he invited her out on a date (there is another reading where "to take out" means to destroy or "take down" but let's ignore that reading here).  Interestingly, (f) cannot be fronted.

I think this is pretty interesting stuff.  I'd be curious to hear which direction you are heading in, Stefan.  Potential things to explore might be 

Tthe more semantically transparent, "adverbial" particles seem to front better than the more resultative, "idiomatic" ones (I use this term loosely because they are all idiomatic to some extent).  I haven't checked the ones above, but separable vs inseparable particles may behave differently too.  But my sense is that this is not the primary distinction that affects the ability to front.

Stefan,

This thread woke me up!. 

Over the last four years I have been compiling a list of Contemporary American English Phrasal Verbs (verbs plus particles) in My Idiolect. At about 600 I knew I was getting close to exhausting the list, but today I have 2563. By collecting such a large database, I have been able to draw some preliminary generalities about the group.

Unlike some more prominent researchers, I do not call particles prepositions and I do not call them adverbs. I feel that there is a differentiation between these, and the way to determine which group a particular instance of these words that have various grammatical uses belongs to, is to perform constituency tests. It turns out that, aside from contrastive stress/movement as mentioned by Benjamin above, if an element is frontable in a sentence, then it fails the test of being a verb particle. So my flat answer of "no" is based on the definition of verb particle.

[* indicates an ungrammatical structure]

Up went the balloon. Up=adverb.

John looked up the pipe (at the mouse). Up=preposition

*John looked the pipe up (at the mouse). Up=preposition

Up the pipe, John looked. / Up the pipe looked John. Up=preposition

*Up John looked the pipe. up=preposition

John looked up the number (in the catalog). Up=verb particle

John looked the number up. Up=verb particle

*Up the number, John looked. / *Up the number looked John. Up=verb particle

*Up John looked the number. Up=verb particle.

I agree with Glenn that  one must distinguish prepositions from particles.  And I broadly agree that one can make a distinction between particles and adverbs.  

However, the question for me is how does one make a principled distinction between particles and adverbs (other than just stipulating it).  For instance, not all adverbs can front (it depends a lot on their function) e.g. *Well he ran.  *Considerably he aged etc.  So the fact that certain particles cannot front doesn't automatically make them not-adverbs.  It just raises the question of which adverbs can front and which ones cannot.

So if we could agree on some way of testing whether something like "up"  is a particle or an adverb, then we could try and see if the distinction correlated with the ability to front or not?

just thoughts

It's good to see we're all excited by the topic of particles, but it seems to me Stefan's question is being ignored here, which was not whether a particle can be distinguished from an adverb (yes it can!) nor whether there's a distinction to be made between "up" used a particle and "up" used as a preposition (of course there is!), nor even about whether it's only spatial particles that can be preposed (in general, yes, but note that certain aspectually used particles can also prepose, e.g. "on and on he sang"). His question was whether transitive verb-particle combinations allow the particle to be preposed. The answer, I think, is that even with spatial particles, this is quite hard but not impossible, given (admittedly rare and recherché) authentic examples like the following:

And in they brought the culprit of thirteen, A boy with bright dark eyes and bright gold hair (www.malespank.net/stories/story-14388.html)

Then off they carried him at a run. (http://www.forgottenbooks.com/readbook_text/The_Downfall_La_Debacle_a_Story_of_the_Horrors_of_War_1000274463/225)

The question, then, is why particle preposing seems more difficult with transitive verbs than with intransitive ones. The answer probably lies in the fact that preposing is a means of putting an element in an unusual position, for a variety of pragmatic reasons. To sum them up for intransitive verb-particle combinations:

1. Contrasting the direction expressed by the particle with other directions, by putting the particle in a focal positon, e.g. I expected the plane to pitch up, but - down it went.  (http://www.rcgroups.com/forums/showthread.php?t=1703530&page=3)

2. Putting the emphasis on the subject by moving it in end-focus position, e.g. Then along came the most beautiful girl our eyes had ever seen. (http://www.cssdesignawards.com/dd/design-stories-6-the-challenging-lives-of-ceos-and-founders/39/)

3. (rare) Proposition confirmation, e.g. she remained in place for about five seconds before she went in, but in she went! (= but go in she did) (http://www.reddit.com/r/Guildwars2/comments/2y1s8v/saw_someone_struggling_at_a_wvw_vista_dropped_a/)

I suppose the reason why preposing works less easily with transitive verbs is that there is then one more constituent in the clause which may vie for the reader's/hearer's attention, while preposing needs all the attention to be directed to the particle (in case 1), to the subject (in case 2) or to the particle as representative of the whole situation (in case 3). 

@Mark

Agree. In the construction He ages considerably the adverb seems to be similar to a resultative, and resultatives of other parts of speech seem to struggle to get to the front of the sentence: (based on an advertising campaign of old)

Raid kills bugs dead! ~ *?Dead, Raid kills bugs!

John appointed Mary captain. ~ *?Captain, John appointed Mary. (The second one survives only as a heavy contrastive: CAPTAIN, John appointed Mary, not queen! and even then, it is awkward at best.

I haven't worked this out, so if I shoot from the hip and miss...well, OK. But the suggestion is that there might be another principle that accounts for your examples. Here's the seed of one idea.

@Bert

In order to address the primary question, some groundwork needed to be laid, so thanks for contributing to that. Others, such as Huddleston and Pullum, would not concede some of what you conceded so quickly. I do.

You have shown that somebody accepts the in they brought/off they carried constructions, but I have trouble accepting those as English sentences. At best they may be poetic, and you know that those poets have a special license to bend or break the rules of grammar. Your examples in 1. and 3. I contend are adverbs, not particles.

I have worked some with the durative aspect markers, however. Your example points to something problematic.

He sang. (no aspectual marker)

He sang on. (durative)

He sang and sang. (durative)

He sang on and on. (perdurative or very durative).

On and on he sang.

*On he sang.

Some others don't work as well:

?On and on he dreamed.

??On and on he hung.

*?On and on the concert dragged.

I would expect the compound form on and on to be a compound particle, such as up and down in He searched up and down = everywhere. I admit however, that it acts like an adverb, and is fully frontable in the case you gave. Of course, this example is not transitive. On the other hand, if we eliminate all the particles as frontable, we eliminate all the particles associated with transitive uses.

@Mark

You suggested that we need to agree on a definition of particle and more specifically verb particle so we can distinguish them from adverbs, prepositions, and the like since the same words--or things that look like the same words--are used among these categories.

I am sure this is kindergarten work for Stefan (and maybe the rest of us, too), but what would serve our purpose? I harvested this group from the internet, fully knowing the danger.

Wikipedia: In grammar, a particle is a function word that must be associated with another word or phrase to impart meaning, i.e., does not have its own lexical definition.

Cambridge Dictionary Online: a word or a part of a word that has a grammatical purpose but often has little or no meaning: In the sentence "I tidied up the room", the adverb "up" is a particle.

grammar.about.com: A word that does not change its form through inflection and does not easily fit into the established system of parts of speech.

Many particles are closely linked to verbs to form multi-word verbs, such as go away. Other particles include to used with an infinitive and not (a negative particle).

Oxford Dictionaries: Grammar. A minor function word that has comparatively little meaning and does not inflect, in particular.

EXAMPLE SENTENCES

(In English) any of the class of words such as in, up, off, over, used with verbs to make phrasal verbs.

Summer Institute of Linguistics (SIL.org): A particle is a word that does not belong to one of the main classes of words is invariable in form, and typically has grammatical or pragmatic meaning.

A verbal particle is a member of a closed class of particles which co-occur with some verbs to form phrasal verbs. In some languages, verbal particles are identical to certain adpositions. Example (English)

The item give up is a verbal particle, as in the following:

        He gave up smoking.

        He gave smoking up.

----------

Of these five definitions of particle, I think the Oxford Dictionaries hits it closest to the mark. My vote for the worst job of defining the term is a tie between SIL and Cambridge Dictionary Online. SIL offers an example of a verbal particle, but erroneously uses the entire phrasal verb—verb plus particle—as the example. Cambridge calls the adverb “up” a particle. So is it an adverb or a particle? If adverbs and particles are the same, then the claim is that adverbs “have little or no meaning.”

All these seem to fail. Need more?

Transitive not separable

*About the disgruntled patron fumed the perfume.

*About we’ll see that.

*Across I ran my kindergarten teacher.

*Across as the fiddler came a fraud.

*After I will lust any woman.

*Against the politician will run his toughest opponent.

*Around the thief danced the subject of when he went home.

*As the clown counted an employee of the circus.

*At the recruit jumped the opportunity.

*Away with she can do a second car.

*Back at Tom got Jeff.

*Behind I always stood a great idea.

*Between my lover would never come me and my dog.

*By you should stand in case Martha arrives late.

*Down on looked the king his valet.

*For the dolt fell the same old trick every time.

*From Ali hails New York.

*In on Joe wanted to get the deal.

*Into the waves slowly ate the beautiful dune.

Transitive, separable

*Off Capone bumped his opponents.

*Off Millie laughed the preposterous charges.

*Off Carol paced 20 meters.

*On the grocer egged his silly sister.

*Out Peter finally asked his neighbor, but she refused again.

*Out the footballer held for more money.

*Over the men paved the parking lot.

*Through the poet struck those inept words.

*To the ambulance driver brought the little girl yesterday.

*Together Mindy and Dave put the party in two days.

*Under I was snowed a week’s worth of old work.

*Up he was brought by a single unemployed father.

*Up he cooked a pan of hot dogs for the fiesta.

*Up they lined quickly for free stuff.

*Up did you scratch the furniture, Logan?

*Up they used all the toilet paper before we arrived.

 Di-transitive

*After the company modeled their sales force WalMart.

*In on Emilio cut me the million Euros.

*In Mom finally caved to us on the vacation plan.

*Around the motley crew fished in the box for some treasure.

*Out hear me on my idea.

Hi Glenn Bingham,

Thanks for your answer!

Well, this does not work. Basically, the who is who of German grammarians claimed that you cannot front particles. Some even defined particle verbs as those verbs that involve a preverb that cannot be fronted. Nevertheless there are cases in which the particle can be fronted.

I wrote about introspection and corpus linguistics here:

https://www.researchgate.net/publication/258340540_Corpora_and_Syntax

One case study is particle verbs. You can find the particle verb fronting data in a lot of other publications as well. For instance here:

https://www.researchgate.net/publication/258432626_Syntax_or_Morphology_German_Particle_Verbs_Revisited

And here in my book about complex predicates (which is open access now):

https://www.researchgate.net/publication/258332361_Complex_Predicates_Verbal_Complexes_Resultative_Constructions_and_Particle_Verbs_in_German

Apart from this Bert already pointed us to his work actually providing examples.

My view on such situations is: We have to check carefully whether there are examples that have the structure we are looking for and if we find any, we can study what the conditions are under which these are possible and in what way they differ from those that are strictly ungrammatical.

Best wishes

Stefan

Chapter Corpora and Syntax

Chapter Syntax or Morphology: German Particle Verbs Revisited

Book Complex Predicates: Verbal Complexes, Resultative Constructi...

 Stefan

From our quick acquaintance, it seems Bert and I think much alike on fundamentals. However, I don't buy into his examples with much enthusiasm, as I judge them differently. There might be some British-American conflict involved in some of these judgements/judgments.

(366) a. They saw the survivors required dry and warm clothing, so off they took them to their own cabins to fit them out with everything they could.

I don't find a strong sense of the phrasal verb surviving in the output. Take off = begin a journey. It is awkward at best. Indeed, the transitive use of this phrasal verb is awkward in itself since it doesn't seem to fit the separable mold: They took the men off to their cabins ~ ?*They took off the men to their cabins. Intransitive: The men took off to their cabins. *Off the men took to their cabins.

b. About ten o’clock there was a flag of truce and in they brought the Captain and three men.

I only read this "in" as an adverb. It does not preserve the meaning of the phrasal verb. Bring in = arrest and deliver to court or jail (place of the authorities).

c. Wood and I boarded the Vixen, and there we got Lieutenant Sharp’s black Cuban pilot, who told us he could take our transport right in to within a few hundred yards of the land. Accordingly, we put him aboard; and in he brought her, gaining at least a mile and a half by the manoeuvre.

This seems to be an adverb in parallel construction to the "right in" above it.

(367) At that moment in was brought a loin of stag and other venison in great plenty, and rich golden plate adorned the table, with great lidded goblets of gold, and magnificent golden candlesticks bearing great candles.

It does not ring of either sense of the phrasal verb that I have: Bring in = arrest and deliver / earn. What is the original phrasal verb supposed to mean?

(368) *At that moment in four waiters brought (or in brought four waiters) a loin of stag and other venison in great plenty…

Same as above. Think of the opposite situation. They brought the venison out. vs. They brought out the full flavor of the venison with the special spices.

The first seems just adverbial, independent of the verb: out = outside, escaping a container. The second is a phrasal verb: bring out = emphasize. I think the venison in your example ends up indoors, but not arrested in some metaphorical sense. If it were hunters rather than waiters...

I anticipate a (felicitous) rejoinder here, but I will not speculate.

Hi Glen,

I didn't judge these examples; I merely reported their occurrence! All I can say is that they are very rare; I don't think I found any in the available linguistic corpora when I studied the phenomenon. But in spite of their rarity, it seems as if you judge the verb-particle combinations in these constructions precisely in the way I understood them: as fully compositional combinations of a spatially used particle and a motion verb. We only differ in our terminology: while you would call the particle an 'adverb' here, I would still speak of a verb-particle combination or a phrasal verb (which I take as synonymous), but that's just because I adopt a broader definition of phrasal verb (namely as also including non-idiomatic combinations). I agree with your comments on these examples: when the particle is preposed, any idiomatic reading that the verb-particle combination in other constructions might have is blocked. So, in a sentence such as

Out they took her

the interpretation can only be 'They took her outside', not 'They shot her'! (I don't think the sentence is incompatable with the idiomatic interpretation 'They took her out on a date', but then the 'on a date' aspect has to be added pragmatically to the literal reading of 'They took her outside (i.e. of her house for example)').

Oh yes, that ?*They took off the men to their cabins sounds strange (to say the least) is simply due to an observation that has been made in the literature, namely that when there's a PP, the particle typically follows the object NP so as to be adjacent to the PP, especially if (as is the case here) the particle and that PP express a single complex trajectory or so-called 'path'. For me, that is not a reason not to consider it a particle anymore, but if you want to call it an adverb, I guess that's fine. And if you think the preposed version ... off they took them to their cabins is rather odd, that's certainly fine with me as well. Our present-day judg(e)ments may differ from the likely acceptance of this sentence when it was first produced, which was over a century ago. (It's taken from Captain Arthur Rostron account of how he rescued passengers from the sunken Titanic, written a year after the tragedy.)

Bert,

Should we take Mark's challenge and try to define particle? Will it separate adverbs from verb particles? Do you adhere to the logic presented above about the Cambridge definition? Do particles and adverbs have the same mixture of content and function?

Stefan, in one of the articles he put forward in this thread, called German ein = 'tin' a particle. But that seems suspicious to me. It seems as if it has all content and no function. Do you agree with that view?

Hi Glenn, "ein" is related to "in". It is used in transparent and non-transparent particle verbs. Its meaning is not "tin". It is when you combine it with "dosen" that it means "to put in a tin" ("Dose" = noun for tin). "dosen" as a verb does not exist. It only comes with "ein" as "eindosen".

Best wishes

Stefan

Thanks for the direction. I will read that again. My Bingemann family left Germany by 1732, and we are a little out of touch with modern German. ;-) I need all the help I can get.

But that is good to hear. It seems that you agree that one of the criteria for being a particle is that it bears little or no "content" meaning.

Hi Glen,

Giving a watertight definition for "particle" might not be easy. I usually recognize one when I see one, but I probably use a whole battery of diagnostics, including obviously whether it belongs to the list of words traditionally recognized to be particles (up, out, off, down, back, forth, etc.). It's true some of these items lead a double life as ('transitive') prepositions (e.g. up, in, ...) and those that don't (e.g. away, back, ...) could perhaps be called adverbs. If the item can be put before the direct object NP, it's definitely a particle, but this test doesn't work with intransitive verb-particle combinations (e.g. sod off) and many transitive verb-particle combinations only occur in the 'split' ordering (e.g. work one's head off).

In any case, what won't work as a criterion is the one that says that the item must be semantically vacuous or at least should express something other than a purely directional meaning. This is because some apparently non-compositional (in the sense of non-spatial) combinations may still be semantically motivated by a conceptual metaphor (even if speakers needn't always activate that metaphor when using the combination), as in the case of figure something out, which might involve the idea of a solution being figuratively locked up inside something. And a particle's semantic emptiness in the case of e.g. tidy up might be an illusion triggered by the fact that the meaning of the particle (e.g. 'resultativity') overlaps with a meaning independently expressed by the verb.

Bert, I might temper your reliance on tradition, having found many things I was taught about language that just aren't true; otherwise, I find nothing to complain about.

If we toss all prepositions, subordinators, verb (or other) particles, and adverbs into one bag and label them relational markers, I think we can go a long way toward sorting them out by the complements they take as suggested in your last message.

R Ø => R = adverb I was there before.

R NP => R = preposition I was there before you.

R CP (clause) => R = subordinator I was there before you arrived.

The problem is that the verb particle can look like an adverb or a preposition complementwise, but must have a critical structural difference: John went before the magistrate. This is ambiguous between the reading that John got lunch in line before the magistrate did and John stood at the bench in court to get the magistrate's ruling. The linear order is the same, but the structure is different. Basically: John went [before the magistrate] (to get lunch). John [went before] the magistrate (to get his sentence). Although I am not yet prepared to be absolutely specific about what in that structure prevents fronting of a particle while allowing it for an adverb or preposition, this is where I think the crucial difference lies in differentiating verb particles, adverbs, and prepositions. And this brings us back to my initial response to Stefan's question: particles do not front by definition.

100 point Bonus: Come on out from up inside of that culvert! List the parts of speech and show the structure of the VP in a tree or in labeled brackets or any systematic notation.

Uh-uh, John went before the magistrate, whatever its interpretation, has the following syntactic structure:

John went [before the magistrate]

In one case (where John went to get lunch before the magistrate did), the verb and the preposition (if that is what it is) do not form an idiomatic combination; in the other case (where John appeared before the magistrate to get his sentence), they do form an idiomatic combination, but that doesn't mean that they form any sort of complex verb, structurally speaking; before is (definitely) a preposition in that case, forming a PP with the following NP.

What's the proof? It's at least two-fold:

1. The verb and the preposition can be separated by a constituent, as shown in the following authentic examples (brackets added), which would be impossible if they formed a complex verb together

he went voluntarily [before the Magistrate] (http://www.oldbaileyonline.org/browse.jsp?path=sessionsPapers%2F18440701.xml)

Meeting a poor man who had been falsely accused, Jesus went with him [before the magistrate] (truthbook.com/urantia/topical-studies/fairness)

2. The preposition and the following NP can jointly form the focus in a split sentence or a wh-question, which shows they form a syntactic unit together:

It was [before the Police Magistrate] that he went.

[Before which Magistrate] did he go?

So, just a because a verb (here: go), when combined with a PP headed by a specific preposition (here: before), adopts a more abstract meaning -- maybe one could imagine going before a Magistrate by skype, so not literally going anywhere! -- this doesn't mean that the verb and that preposition form a syntactic unit together. 

As for the 100 bonus points: see my publication on this sort of complex structures:

https://www.researchgate.net/publication/228656614_The_grammar_of_complex_particle_phrases_in_English

Basically, in Come on out from up inside of that culvert!

on is a particle specifying the particle out, which is the head of a particle phrase taking the PP from up inside of that culvert as a kind of complement whose head is from. From takes up inside of that culvert (which as a whole could be replaced by it or there) as its complement; this complement in turn is a PP headed by the two-word preposition inside of, which is specified by the particle up and which is complemented by the NP that culvert. I'm happy with 90 of the bonus points.

Article The grammar of complex particle phrases in English

Bert: Bonus--You scored higher than I did and I made up the problem!

Of course the ambiguity was from before being used temporally and spacially. You are so easy to get along with, I could not force out many details. I finally goaded you into offering some constituency tests and attacking the problem. However, your publication answers a lot of my questions. As it turns out, much of our fundamental stance coincides. What I call adverbs and particles you call particles. I guess the burden is on me to show that both categories are needed.