Author Topic: Relevance Theory and processing effort in proper nouns with articles  (Read 745 times)

Offline Matt Longhorn

  • Jr. Linguist
  • **
  • Posts: 17
I am currently looking at the use of the Greek article ho in Hellenistic Greek.

There have been a lot of discussions in the literature about its use and non-use in the New Testament and how it effects the salience of elements in the discourse.
Some authors have applied a general rule that a (proper) noun may be introduced in the discourse without an article and then subsequent uses will use it. In order to modify discourse prominence, the absence of the article with a proper noun still activated in the discourse is said to add salience to that referent. I am wondering whether there has been any work done on processing effort involved in the use / non-use of (definite) articles associated with proper nouns?

From a relevance theory perspective it seems that it would be common and reasonable to introduce a discourse participant without the article. Using Christopher Lucas' terminology based on Hawkins' work, that entity is then capable of being represented as a mutually manifest P-Set which explains the use of the article with it subsequently. Lucas' approach also explains instances where an entity is introduced with the article and no prior reference.

In many switches between two activated discourse participants a post-positive marker of  narrative progression (de) is preceded by the article - (ho de), the article thus being used almost pronominally. It seems to me that this common device from an RT perspective must be due to a desire to reduce processing effort in decoding extra linguistic information that would be involved if the proper noun was used each time.

In some instances the switch back to an already activated participant in the discourse will be done with a proper noun and an article; from an RT perspective I would presume that this would thus highlight the salience of that participant by incurring an extra processing cost.
Greek can also switch to an already active discourse participant with a full noun phrase without the article. Again, I am wondering whether this would incur extra processing due to the lack of signal that the article provides that this is a member of an already activated P-Set in the hearer's mental representation. Without the article, the hearer would thus need to seek to assign a reference to the concept triggered by the proper noun and it would be the same. Asking the hearer to go to this effort is mitigated I am thinking by the fact that they are highly accessible to the hearer and therefore it is not an unreasonable device to use.
Again, I am wondering all of this from an experimental pragmatics perspective, namely whether there is any work on neural activation across the use of articles with discourse participants or other tests of processing effort.

I am pretty new to linguistics so apologies if my terminology is all over the place. I have also horrendously over-simplified the literature on the article... literally books have been written about its use and non-use and I can't accurately represent it all here

Offline Daniel

  • Administrator
  • Experienced Linguist
  • *****
  • Posts: 1796
  • Country: us
    • English
Re: Relevance Theory and processing effort in proper nouns with articles
« Reply #1 on: June 09, 2018, 07:51:49 PM »
Quote
Some authors have applied a general rule that a (proper) noun may be introduced in the discourse without an article and then subsequent uses will use it. In order to modify discourse prominence, the absence of the article with a proper noun still activated in the discourse is said to add salience to that referent.
That's interesting, and makes some sense because definiteness is at least sometimes anaphoric in discourse, e.g., "defined" (rather than "specific" or other types of "definiteness").
I have not previously heard of this particular pattern, but I haven't studied definiteness myself, just seen some things in passing.
Quote
I am wondering whether there has been any work done on processing effort involved in the use / non-use of (definite) articles associated with proper nouns?
I'm not sure about that. It sounds like an interesting study to do.

Someone whose work you might want to consult is Tania Ionin:
https://linguistics.illinois.edu/directory/profile/tionin
She has worked extensively on definiteness from a psycholinguistic perspective. Most of her work is about language acquisition, but in doing so she also looks at native speaker control groups and has contributed directly to our understanding of what definiteness is semantically. But I don't think she has (yet!?) looked at definiteness combining with proper names. That would be an interesting extension and might fit within the same experimental methodology she has used.

I'm not sure of exactly what language would be best to test with (that's usually a limitation for psycholinguistic work), but there's probably one out there, if you can access speakers of course.

Quote
From a relevance theory perspective it seems that it would be common and reasonable to introduce a discourse participant without the article. Using Christopher Lucas' terminology based on Hawkins' work, that entity is then capable of being represented as a mutually manifest P-Set which explains the use of the article with it subsequently. Lucas' approach also explains instances where an entity is introduced with the article and no prior reference.
I don't think that's wrong, but is it necessary? That's what definite literally means-- previously introduced (defined) in the discourse, right?
But maybe you'll be able to explain that idea of definiteness itself (just be careful that it can mean several different similar things, like specificity or uniqueness).

Also, going back to the earlier Maxims (which can usually be readily derived from Relevance nowadays), wouldn't the Maxim of Quantity suggest that having an extra word in repeated usage would be odd? Then again, maybe it's there to add the additional information that it's the same entity already expressed.
(But compare this to the 'Switch Reference' phenomenon in 'pro-drop' languages, where a subject pronoun is substantially more likely in an introduction instance than in a repeated usage instance, where usually the agreeing verb alone is enough. [This has little to do with the other/older sense of "Switch-Reference" related to verbal morphology, except at the discourse level of same/different subjects.])

Quote
In many switches between two activated discourse participants a post-positive marker of  narrative progression (de) is preceded by the article - (ho de), the article thus being used almost pronominally. It seems to me that this common device from an RT perspective must be due to a desire to reduce processing effort in decoding extra linguistic information that would be involved if the proper noun was used each time.
Hm, see above. Without exception as far as I know, languages tend to drop repeated pronouns rather than use them more often when repeated.

Whether this is an exception to that or something else depends on how you analyze it, but that literature would be worth looking into. I have a fairly substantial bibliography on it I could send you if you're interested. It's not hard to find papers that mention it, though. Cameron's (etc.) work is relevant, as well as too many other papers to name. It is primarily about Spanish and other Romance languages but similar results have been reported elsewhere for typologically distinct languages, like Japanese, too.

Quote
from an RT perspective I would presume that this would thus highlight the salience of that participant by incurring an extra processing cost.
Again, puzzling! Because you could predict the exact opposite for the reasons above.

"Salience" I suppose could be considered in several different ways, including emphasis, but why would that be the case on repeated usage?

Quote
I am pretty new to linguistics so apologies if my terminology is all over the place. I have also horrendously over-simplified the literature on the article... literally books have been written about its use and non-use and I can't accurately represent it all here
It is very interesting! I hope some of my comments might help.

--
Updated to add: I meant to also ask whether or how this pattern fits into the typical grammaticalization pathway for articles, first demonstratives, then definiteness markers, then noun markers (including on proper nouns), etc.
https://en.wikipedia.org/wiki/Article_(grammar)#Evolution
http://www.oxfordhandbooks.com/view/10.1093/oxfordhb/9780199586783.001.0001/oxfordhb-9780199586783-e-42
Author's draft PDF of that: http://www.lattice.cnrs.fr/IMG/pdf/Handbook.gramma.2010.pdf
« Last Edit: June 09, 2018, 11:33:52 PM by Daniel »
Welcome to Linguist Forum! If you have any questions, please ask.

Offline Matt Longhorn

  • Jr. Linguist
  • **
  • Posts: 17
Re: Relevance Theory and processing effort in proper nouns with articles
« Reply #2 on: June 12, 2018, 04:01:37 AM »
Hi Daniel, thanks for the response. I was wondering whether I had been so incomprehensible that no-one was quite sure what to say!
I agree with the following
Quote
definiteness is at least sometimes anaphoric in discourse
. I am considering looking through one of the smaller biblical books with Matsui's work on bridging relations at some point. This may be able to explain a number of instances with initial arthrous proper noun.

Quote
Someone whose work you might want to consult is Tania Ionin
Thanks

Quote
I don't think that's wrong, but is it necessary? That's what definite literally means-- previously introduced (defined) in the discourse, right?
. Not being a linguist and only being exposed to a small subset of the material I would disagree with this. Definiteness through bridging relations, situational context etc would indicate that previous discourse mention or even activation need not be present for something to be definite. Russell's theory of uniqueness of entity has come under a fair bit of flack by relevance theorists. Powell is an interesting read on this. Happy to modify my views though if you have more thoughts?

Quote
(But compare this to the 'Switch Reference' phenomenon in 'pro-drop' languages, where a subject pronoun is substantially more likely in an introduction instance than in a repeated usage instance, where usually the agreeing verb alone is enough. [This has little to do with the other/older sense of "Switch-Reference" related to verbal morphology, except at the discourse level of same/different subjects.])
Urmmmmm, I am well and truly out of my depth here, but will look this up.

Quote
Whether this is an exception to that or something else depends on how you analyze it, but that literature would be worth looking into. I have a fairly substantial bibliography on it I could send you if you're interested.
If you could send that bibliography over it would be really appreciated. I will see how much money I have at the end of the month and see whether I need to break my bank again.

If you are interested, the below is a more fully fleshed out summary of my thoguhts for this personal study that I posted on a biblical greek board. Thanks for all the comments!

I am considering the use of the definite article in the New Testament. Recently I have been listening to and reading a lot of journal articles as well as reading a number of chapters in books on definiteness and reference. I have also been reading about the use of the definite article in English from a relevance theory procedural meaning perspective.

I have found that a number of treatments of the article in the New Testament don’t really deal with the question of what it means to be definite, I see this as a slight missing. Similarly, many treatments of the article are based on intuition rather than experimental pragmatics, these often deal with default means of reference. Whilst not fully disagreeing that there are default ways of referring, I want to see what underlies these, namely why they became defaults.

I clearly haven’t read everything on the subject, however I have read lots of journal articles, chapters of a number of books on NT linguistics, as well as Wallace’s book and Ronald D. Peters’ PHD dissertation. This is on top of all the grammars I could lay hands on.

I am interested in looking at applying Christopher Lucas’ modification of Hawkins’ work on the use of the definite article (see link below) to the New Testament and also Matsui’s work on bridging in understanding the use of the article. I want to also work through papers on experimental pragmatics looking at processing effort in reference assignment.

https://www.academia.edu/2129603/Defini ... ommodation

My current suppositions with regards to participant reference specifically, for the most part agree with the outcome of other works that I have read on NT linguistics. The difference being the relevance-theoretic and hopefully experimental pragmatics underpinning.

Initial introduction of participant in narrative with no article unless either:
1. Activated in the discourse already through spreading activation or some other mechanism e.g. bridging
2. Accessible as a mutually manifest P-set
Subsequent reference through articles or pronouns. Some relatively recent work on processing cost of discourse reference has shown pronouns to be more costly than nouns therefore impacting relevance

Where switching between previously activated participants in an embedded discourse section
1 The use of ho de is used as a means of reducing the processing cost that would have been involved in the use of full noun phrases. The article anaphorically marking the noun phrase used as a mutually manifest P-Set. The cognitive principle of relevance will allow the hearer to stop at the most accessible reference assignment which will usually be that one in the recent discourse
2 The use of the article with a noun phrase to predominantly create an expectation of more cognitive effects where the previous full reference is relatively recent. This additional expectation is due to the increase in processing effort due to decoding the greater lexical information. Unlike previous work that I have looked at I would also like to look at this use sometimes perhaps being simply to help the hearer keep track of which participants are involved thus the increase of relevance through reduction in processing cost in some instances. This would divert from perspectives that see this as necessarily highlighting the salience of the information following and would need to look at the length of the reported discourse interactions.
3 The use of the anarthrous noun phrase to increase expectations of relevance. This would be through the lack of indication that the noun phrase is a mutually manifest referent, thus requiring reference assignment by the hearer. Given that this is a recent referent then this would be highly accessible to the hearer and thus not require an undue amount of processing.

Rather than just applying the relevance theory principles intuitively I would be interested in researching this from an experimental pragmatics perspective using works on modern languages as an analogue.

In particular:
1 Processing effort based on use of articular vs anarthrous proper nouns
2 Participant reference in participant switching

Offline Daniel

  • Administrator
  • Experienced Linguist
  • *****
  • Posts: 1796
  • Country: us
    • English
Re: Relevance Theory and processing effort in proper nouns with articles
« Reply #3 on: June 14, 2018, 05:51:44 AM »
I'm glad my response was helpful.

Quote
Quote
That's what definite literally means-- previously introduced (defined) in the discourse, right?

. Not being a linguist and only being exposed to a small subset of the material I would disagree with this.
I was referring to the etymology of the term: definite means defined, as in established or known from the discourse context. Of course now it's just a technical term not necessarily always meaning (only?) that.

Quote
Definiteness through bridging relations, situational context etc would indicate that previous discourse mention or even activation need not be present for something to be definite. Russell's theory of uniqueness of entity has come under a fair bit of flack by relevance theorists. Powell is an interesting read on this. Happy to modify my views though if you have more thoughts?
There are (at least) two types of definiteness that interact, and can't necessarily be substituted or explained by the other.
1. Established in context (literally "defined"): "I bought a boat. ... then the boat sank."
2. Specificity/uniqueness: "the book on the table", suggesting either there is only one, or it is the obvious one.
The first may sometimes have a sort of anaphoric ('referring back', as in pronouns) sense.

Quote
Quote
[Switch-Reference]
Urmmmmm, I am well and truly out of my depth here, but will look this up.
The term "Switch-Reference" is used to refer to when there is a change in subject from one clause to the same, in contrast to maintaining the same subject:
1. He arrived home. He watched a movie. SAME SUBJECT
2. He arrived home. She watched a movie. DIFFERENT SUBJECT

Different languages have different ways of expressing this.

In some languages, there are different endings on the verbs indicating whether the subject of a following clause is the same or different:
https://en.wikipedia.org/wiki/Switch-reference


But that's getting away from the main point I wanted to make:

In languages where subjects are optional (these are called "pro-drop languages" by linguists, sort of an odd term, meaning "pronoun-dropping"), then:
1. When many verbs have the same subject, it is normal to omit repeated pronouns.
2. When pronouns are used, very often they refer to new (different) subjects.
So this means that in general, the use of pronouns is associated with change of reference in the discourse.

Note that this is a somewhat similar situation to what you describe for Greek, but exactly the opposite tendency: pronouns tend to disappear in repeated usage, while you find repeated usage of "ho" when the same referent is mentioned repeatedly.

For the relevant sense related to the use of pronouns, see this reference:
Cameron, Richard. 1995. The scope and limits of switch reference as a constraint on pronominal subject expression. Hispanic LinguisLcs 6/7. 1–27.
And look up various papers citing that one (it's probably the most cited article, though there are many more if you want to read them).

---

Good luck with your research. The proposal sounds interesting, especially investigating those ideas experimentally.

The potential weakness in the argument regarding Relevance Theory, and not necessarily specific to your work, is that it seems equally plausible to come up with the opposite hypothesis for similar reasons also motivated by relevance. This is why I referred to omission of pronouns in same subject contexts above.

Roughly, if the speaker and hearer already know about the subject, then why would we need to reinforce it?

You say the repeated usage somehow highlights its salience, and I can see that. But wouldn't this mean that you could get a contrast between repeated use of salient subjects, versus repeated use of unimportant subjects? Or, if not that, then try to find a way to distinguish these hypotheses.

The experimental results may be helpful in this, depending on how you design the experiment!
Welcome to Linguist Forum! If you have any questions, please ask.

Offline Matt Longhorn

  • Jr. Linguist
  • **
  • Posts: 17
Re: Relevance Theory and processing effort in proper nouns with articles
« Reply #4 on: June 18, 2018, 05:18:46 AM »
Thanks again Daniel, useful clarifications and some good directions for me to look. I also just saw the link to your thesis - I will download that later and get some time to read through.
I won't be doing any experimentation myself, far too incompetent and under-educated for that! I am just hoping to get a view of experiments done in other languages that i can then try to apply any consistent principles back to Greek. Studies on modern Greek obviously, where I can find them, would be a good place to start!
Quote
Note that this is a somewhat similar situation to what you describe for Greek, but exactly the opposite tendency: pronouns tend to disappear in repeated usage, while you find repeated usage of "ho" when the same referent is mentioned repeatedly.
There is the option to use pronouns also in these constructions, the pronominal use of the article is almost a short hand. I have deliberately avoided mentioning them below trying to avoid complicating things even more. Perhaps I will need to increase my scope to look at pronoun use... in attempting to focus on something really small narrowly defined, as normal I seem to be multiplying stuff I need to read!

Quote
The potential weakness in the argument regarding Relevance Theory, and not necessarily specific to your work, is that it seems equally plausible to come up with the opposite hypothesis for similar reasons also motivated by relevance. This is why I referred to omission of pronouns in same subject contexts above.

Roughly, if the speaker and hearer already know about the subject, then why would we need to reinforce it?
My thoughts on reinforcing the subject when it is already known by speaker and hearer are pretty ill-thought through but would basically be as follows. In some reported dialogues, where there is a lot of switching between participants, surely there comes a time that anchoring the current discourse back to a named referent can help the comprehension. If half a page of text contains 10 switches between short interactions, then using the name of the current speaker might be helpful even if not really flagging the material as important. I would want to look into the possibility that some instances considered as marked reference in some grammars may just be the author needing to help the audience keep track of who is speaking. As I said - not carefully thought through, just something nagging at the back of my mind.

Quote
You say the repeated usage somehow highlights its salience, and I can see that. But wouldn't this mean that you could get a contrast between repeated use of salient subjects, versus repeated use of unimportant subjects? Or, if not that, then try to find a way to distinguish these hypotheses
I will have to think about this. Thanks for pointing it out.

Thanks again

Matt