Disagreement Results

  • Influence on Listener Kappa Insignificant
  • Influence on Speaker Kappa Insignificant
  • Abandoned Kappa 0.61, PA 0.92
  • Agreement Kappa 0.21, PA 0.72
  • Answer Kappa 0.53, PA 0.86
  • Conventional Kappa Insignificant
  • Exclamation Kappa 0.37, PA 0.98, Kappa Significance .1
  • Performative perfect
  • Info-level Kappa 0.41, PA 0.89
  • Other Forward Function Kappa Insignificant
  • Response-to Kappa 0.46, PA 0.66
  • Self-talk Kappa Insignificant
  • Statement Kappa 0.55, PA 0.75
  • Understanding Kappa 0.3, PA 0.72
  • Unintelligible Kappa Insignificant
  • Info-request Kappa 0.62, PA 0.90

    Comments on Map Task Dialog d204

    General Comments

    email mcore@cs.rochester.edu to add your comments.

    A's comments: what's the forward function of Okay or of repetitions that indicate agreement? like utt15, utt45. can't be commit (no action)

    Editor's response: Does there have to be a forward function? In this case, I'd say no.

    Comments on Individual Utterances

      1     G: Okay. 
      B - don't know what happened before, can't label
        (note this is start of dialog, nothing happened before)
      2       Do you have a start X? 
      3     F: [Yeah](0) 
      4     G: [up top?](0) 
      5     F: [Up](0) by sandy shore. 
      6     G: By sandy shore. 
      Y - I think this utterance was just to provide grounding, and isn't 
           a reassert  according to the criteria in the manual (I think it 
           fails the true/false test).
      A-is it a reassert? or other forward function?  36 is reassert
      7       Just below that do you have a well 
      8     F: No. 
      9     G: Oh. 
      C - could be self-talk
      10       Great. 
      Y - This one is interesting.  If we equate this utterance with 
          "That's great." then it can be responded to with "No it isn't.", 
          and so by our criteria it is a statement.  But I don't think the 
          speaker is trying to change the belief of the addressee - 
          he's just giving a reaction/opinion.  It's also interesting that
          he means the opposite of what he says.
      Z - Comment on the state of the task?
      B - this is an assessment; i'm not sure how to mark its backwards function
      A=can it be considered an assert? like if it were 'that's bad'
      11       Do you have any land marks just below 
      12     F: Uh no. 
      13       About half about a quar- th- 
      A=not sure about abandoned. 'about' is used to understand utt14
      14       third of the way down I have some hills. 
      Y - I didn't tag this as an answer to utt11 since utt11 is a yes/no 
            question, a direct answer was given in utt12.
      15     G: Okay. 
      A=which forw-function? could it be Reassert?
      16     F: And I got a uh 
      17       local resident 
      D - Intonation indicates for utt16 to 18 that these utterances belong to the 
             same communicative function. For this reason they are tagged 
            identically here.
      18       and an iron bridge. 
      19     G: An iron bridge. 
      Y - Seems to be repeating to show that he "heard" what was said
            but his intonation indicates that he hasn't accepted the 
            information, and perhaps that he'd like elaboration.
      Z - There's something else in the intonation.  He doesn't see it.
      C - he understands "iron bridge", but to put "repeat-rephrase" would
       be misleading --- this signals a problem.
      A=could it be other-stat, instead of other-FF? not reassert because G doesn't 
          have bridge (cant hear sound for utt19)
      20     F: In the middle of the map 
      21       right at the top 
      22     G: Okay 
      D - forward function is maintaining the communication, why is okay separated 
           here and not in 28?
      C - not really separate enough intonationally that I like coding ss 
      23       I don't have the iron bridge. 
      24     F: Do you got the [My](1) 
      C - utt 24 not abandoned --- just stumbles when interrupted
      25     G: [Do](1) y- 
      26     F: You got the local resident? 
      D -utt24 and utt 26 parts of the same speech act
      27     G: Yes I do. 
      28     F: Okay, my iron bridge is right above that 
      Y - This could be separated into 2 utterances
      29     G: [Oh, okay](2) 
      D - forward function maintaining communication
      A=usual problem with Okay (as in 15, 45)
      30     F: [like it's](2) like uh ... 
      A=not sure how to deal with 30-33. 30-31 dont seem to contribute to dialogue, 
         32-33 yes
      31       [going across](3) 
      32     G: [It crosses](3) 
      K-utts 32 and 33 are treated as a single utt; also, this is really
          more of a simultaneous "chant" than a repetition acceptance
      A=utt32-utt33 together. don't think they're abandoned, as F says 
                 yeah in 35
      33       [the bay](4) 
      34     F: [go-](4) 
      D - interpreted as part of 36 and tagged identically,
      35       Yeah 
      D - not considered to be part of segment 22,  F starts  his description in 34, 
             answers the question in 35 and restarts his description, then s
      C - simultaneously half answers, but he didn't bother understanding 
            question, I think --- just brushed off G
      A=like Okay's
      36       Going across the the river there. 
      Y - There is actually a little correction going on here -
           "bay " to "river"
      37       Like 
      Y -  utt37-utt40 should be joined. 
           I wasn't sure how to tag this.  He's just refering to an area on
            the map, he hasn't said anything about it.  Does this actually 
            "make a statement about the world?"
      C - don't know why says this; discourse marker?
      D - can't see how to tag 37 separate from 38, 39 from 40 and 41 from 42
           seems different from "take the people to delta"
      38       below the local residents 
      39       about an inch 
      40       to the left 
      41     G: Is 
      42       the hills? 
      Y - He's guessing at what the other participant
          was going to say about the map location refered to.
      K-utts 41-42 treated as single utt. Question implies presence of hills
      43     F: I got the h- 
      44       That's where my hills are 
      A=not sure whether it's answer. intonation doesn't indicate answer
      C - ignores proferred completion
      45     G: Okay. 
      E - i called this task because in the maptask domain, an ok implies
           that Follower has it too)
      A=as usual, what's forward function of "okay"
      46     F: Um to the 
      D - because of restart in utt47, neither a forward nor backward function is 
            seen in 46
      47       if you go right of the iron bridge 
      48       I have a woodland 
      49     G: See Okay. 
      A=FF as in 45 etc
      50     F: [Okay](5) 
      Z - This one sounded like a cue phrase to me.
      51     G: [I have a](5) 
      52     F: if you follow 
      53       if you follow the brook down 
      Y - utt53 and utt54 should be joined.  Again, he's refering to a map
         location, but hasn't actually made an assertion yet.
      54       [the babbling brook](6) 
      55     G: [The forked stream](6) 
      C - sort of mini-misunderstanding correction
      56     F: The babbling brook 
      57       or whatever 
      A=??? seems to accept different descr of stream in 55
      58       I have a dead tree 
      59       Do you have that 
      60     G: I have the dead tree at the fork. 
      61     F: Yeah 
      A=without rest of dialogue difficult to understand what utt61 is doing

    Last change: 13 March 98 by mcore

    Click here to send in comments or questions.

    Back to the Dialog Annotation Page