"Intelligent transportation communication systems"?

« previous post | next post »

This morning's email brought an invitation to contribute to a "Special Issue on Intelligent Transportation Communication Systems" (for this journal). It took me a little while to figure out that conversing with cars (which I'm definitely in favor of) was not what they had in mind. And this process  reminded me of how difficult it can be for humans — never mind machines — to figure out how to parse complex nominals in English. (See "The Stress and Structure of Modified Noun Phrases in English" for some antique thoughts on the subject…)

It's easy enough to figure out that "stone traffic barrier" is probably a traffic barrier made of stone, rather than a barrier against stone traffic — thus [stone [traffic barrier]] rather than [[stone traffic] barrier]. And it's obvious that "rush hour traffic" is traffic during rush hour, and not hour-traffic having something to do with rush — thus [[rush hour] traffic] and not [rush [hour traffic]].

So we're tempted to generalize such observations to the conclusion that all complex nominals in English are built up recursively out of binary pieces.

But what about those "Intelligent transportation communication systems"?

The number of possible binary trees with n leaves is the (n-1)th Catalan number, where the nth Catalan number is

$$\dfrac{(2n)!}{(n+1)!n!}$$

So for a four-leaf expression like "Intelligent transportation communication systems", we get a manageable number of possible trees, namely 5:

This offers a choice among:

  1. [[[Intelligent transportation] communication] systems], meaning something like "systems for communication in intelligent transportation".
  2. [[Intelligent [transportation communication]] systems], meaning something like "systems for intelligent communication in transportation".
  3. [[Intelligent transportation] [communication systems]], meaning something like "communication systems for intelligent transportation".
  4. [Intelligent [[transportation communication] systems]], meaning something like "intelligent systems for communication in transportation".
  5. [Intelligent [transportation [communication systems]]], meaning something like "intelligent communication systems for transportation".

Which is it?  Frankly, I'm not sure. Here's the list of suggested topics, to help you decide:

  • Communications and networking for connecting vehicles
  • V2V, vehicle-to-infrastructure (V2I) and vehicle-to-X (V2X) communications and networking
  • Spectrum assignment and EMC regulations for connected vehicles
  • Wireless technologies for connecting vehicles (propagation models, RF technologies, antenna design, physical layer design, etc.)
  • Link and network layer protocols for connecting vehicles (MAC, routing, mobility management, geo-networking)
  • Multimedia applications for connected vehicles
  • Architectures, and protocols for data dissemination, processing, and aggregation for connected vehicles
  • Networked information processing, decision making, and intelligent control
  • Safe driving and accident avoidance
  • Railway communications and networking
  • Applications and services with connected vehicles
  • Security, privacy, and dependability for connected vehicles
  • Cellular networks cohesion with connected vehicles for various safety and non-safety applications
  • Simulation and performance evaluation techniques for intelligent transportation communications
  • Results from experimental systems, testbeds, and pilot studies for intelligent transportation communications
  • Assessment of impact of connected vehicles on safety, accident prevention, hazard reduction, intersection assistance, cooperative awareness services, misbehavior detection and mitigation services
  • Intelligent transportation, vehicle traffic modeling, decentralized congestion control, highway automation and platooning, autonomous cooperative driving, location-dependent services

I'm still not sure. Since the three-word phrase "intelligent transportation communications" occurs twice in that list of topics, and "intelligent transportation" occurs by itself once, I'm inclined to accept those groupings, which would give us the left-branching option #1:

[[[Intelligent transportation] communication] systems]

But does option #3

[[Intelligent transportation] [communication systems]]

really mean something incompatible with this? Or for that matter, are any of the others ruled out?

In cases like this, we have the impression that we've understood the phrase without having a clear idea of how to parse it. This is partly because the default interpretation of an English noun phrase of the form [X Y] is is pretty vague —  a Y that has something to do with X — and partly because it's often not clear from the context what is really modifying (or otherwise combining with) what.

It was apparently not phrases like this that Gertrude Stein had in mind when she wrote about the excitement of diagramming sentences:

When you are at school and learn grammar grammar is very exciting. I really do not know that anything has ever been more exciting than diagramming sentences. I suppose other things may be more exciting to others when they are at school but to me undoubtedly when I was at school the really completely exciting thing was diagramming sentences and that has been to me ever since the one thing hat has been completely exciting and completely completing. I like the feeling the everlasting feeling of sentences as they diagram themselves.

In that way one is completely possessing something and incidentally one’s self. Now in that diagraming of the sentences of course there are articles and prepositions and as I say there are nouns but nouns as I say even by definition are completely not interesting, the same thing is true of adjectives. Adjectives are not really and truly interesting. In a way anybody can know always has known that, because after all adjectives effect nouns and as nouns are not really interesting the thing that effects a not too interesting thing is of necessity not interesting.

Anyhow, this is why the original Penn Treebank made a decision not even to try to annotate such structure — thus from the "Bracketing Guidelines for Treebank II Style Penn Treebank Project", 1995:

The interpretation of modi fiers that are themselves nominal tends to be highly ambiguous and subject to individual interpretation. For example, in the noun phrase the primary college entrance examination, one person may have a clear intuition that the college is primary, while another may be sure that the examination is primary. Similarly, in U.S. patent and copyright owners , one person may think that the owners are U.S., while another may believe that the patents and copyrights are U.S.

In order to avoid spending large amounts of time imposing arbitrary solutions to this problem, we try to avoid showing any structure for nominal modi ers:

(NP the primary college entrance examination)
(NP U.S. patent and copyright owners) 
(NP the loan and real estate reserves)

In general, we avoid showing either the internal structure or the extent of modi fication of noun modi fiers, regardless of the strength of the annotator's intuition in a particular example.

[But see Vadas, David, and James R. Curran. "Parsing noun phrases in the Penn Treebank." Computational Linguistics 37, no. 4 (2011): 753-809.]

 



10 Comments

  1. Tom Wilson said,

    May 26, 2017 @ 10:43 am

    What if "transportation communication systems" is already a term of art within the circle of people to whom the whole thing is addressed?

    [(myl) Then there's a term-of-art fight between "transportation communication systems" and "intelligent transportation communications". Outcome uncertain:


    ]

  2. Luke Ritchie said,

    May 26, 2017 @ 11:38 am

    Hey! an academic field I actually know something about on Language Log. Your analysis is correct, and something like sense #1 or #3 is intended. Think "communication technology that enables stuff like self-driving cars". Is there really any difference between these and the other options? I would say that parsing "intelligent" as modifying "communications" would describe something incompatible, since these systems are making use of "dumb networks" (which I think may qualify as a term of art) .

  3. Sniffnoy said,

    May 26, 2017 @ 11:42 am

    I think there's a bit of associativity going on here. It's not clear to me that #1 and #3 actually mean different things, nor #4 and #5.

  4. Gregory Kusnick said,

    May 26, 2017 @ 1:00 pm

    From the title, my guess was going to be "systems for getting smart cars to talk to each other", and from the list of topics that appears to be correct. So I'm going with tree #3.

  5. John Roth said,

    May 26, 2017 @ 7:00 pm

    I would go with [intelligent [transportation communication systems]], simply on the basis that there's a lot of concern that [dumb [transportation communication systems]] could cause a lot of havoc (possibly fatal).

    [intelligent transportation [communication systems]] is also a possibility since they basically mean the same thing: a system that operates with some degree of actual intelligence.

  6. maidhc said,

    May 26, 2017 @ 9:17 pm

    I don't think "system" adds much to the meaning so you could leave it out, giving either I(TC) or (IT)C.

    If it's I(TC), then that means there must be such a thing as "stupid transportation communications". I would take that to include transponders on trains or planes, some of which can be pretty simple, but could well be made more intelligent.

    If it's (IT)C, then the question is could transportation be intelligent without being able to communicate? I think it could, possibly. Imagine a robot that taught itself to find its way around a supermarket. When it's at its docking station you could type in "tomato soup" and it would trundle off on its own and bring back the soup without needing to communicate with anything else.

    The list of topics seems to lean to the first interpretation.

  7. Gregory Kusnick said,

    May 26, 2017 @ 9:56 pm

    The suggested topics are heavily slanted toward systems engineering, so from that perspective the word "systems" is not irrelevant; on the contrary, it seems to be the main point.

  8. maidhc said,

    May 27, 2017 @ 2:11 am

    I'm not really sure there is such a thing as systems engineering as a stand-alone topic. A system is a thing or collection of things that accomplishes some task. To me it's the task that is the important part. A communication system is something that implements communication. To me it seems that the topics are slanted towards communication, so my example of the self-navigating robot would not qualify. You can't have communication without a communication system, obviously. But the thing that the papers would be discussing is how the communication is implemented. How cars could talk to each other or to a navigation hub, and the like.

    There are people who would argue that the study of systems is a study in itself. That group would include a younger version of myself. In years gone by I applied for some grants based on that idea, but I didn't get them. So now I have adopted the alternative viewpoint. Call it sour grapes, perhaps.

    On the other side, count in the Rastafarians with their concept of "Babylon system".

  9. Mark Liberman said,

    May 27, 2017 @ 7:01 am

    @maidhc: I’m not really sure there is such a thing as systems engineering as a stand-alone topic.

    This will come as unwelcome news to my colleagues in the Electrical and Systems Engineering Department, formed a few years ago by a merger of the Electrical Engineering Department and the Systems Engineering Department. Someone also needs to inform Wikipedia that their article on Systems Engineering is in need of updating or perhaps deletion…]

  10. maidhc said,

    May 28, 2017 @ 3:21 pm

    The problem with that is that there are many types of systems that are not included in what is called Systems Engineering. My comment was in reference to all types of systems, not that subset that is studied in Systems Engineering. That subset does form a coherent body of knowledge, but I think it is poorly named.

RSS feed for comments on this post