GTRP Dialect transcriptions 1980-1995

How to go about

Nederlands | on 'synthesis' | on saving maps | assignment | per place | © some copyrights

1) We start with a simple example to illustrate the main steps towards a map.
2) Next follow some reminders for 'advanced users'.

App works fine with Windows Explorer 6 (slow), Netscape 7.1 (fast!); with MacOS Safari, Explorer 5.2, Netscape 7.1 and Mozilla 1.7 (fast!)

Reference to the (Dutch) page with transcriptions may also be made by


Last century two well-known Dutch dialectologists (A.R. Hol and A. Weijnen) published maps on -n in word endings. Two smaller versions are given here and here. Add your own version now and find out whether differences exist over time.

It is recommended to have read all introductions on the data and atlas before continuing.

Step 1:  Let us choose the category of 'infinitives'.

Step 2: A list will be shown of all verbs recorded. Choose 'geven' (give).

Step 3: You confirm your choice.

Pre step 4: A list will be shown of all (simplified) transcriptions for each of the dialects.

With other queries than in this example it is highly recommended to quickly scan the list of transcriptions first in order to get ideas on the specific types of variation regarding the current item. This will help formulate your queries more adequate.

At the moment all we want to know is whether an n exists at the end of the infinitive. The query can be straightforward since no other n is expected in g-e-v-e-n .

Step 4: Click the button 'create map' and wait for the query form. Enter n in the very first field, click the 'quick map'-button once and wait for the map to appear.

Just half a minute later we are able to find out that the overall pattern matches that of Hol and Weijnen. Yet it becomes also clear that more -n is spoken in the central area than was indicated on the previous maps. Will this be a new development? Well, it might have been but in this particular case of word final -n the data behind the previous maps appears not to be adequate because of the type of questions in the questionnaire the maps were based on and the number of places the (written) questionnaire was returned from. As regards the type of questions it may very well be that the questionnaire designers (which were not the mentioned Hol and Weijnen) were not particularly interested in word final -n.

Whether we uncovered a new development or not, it certainly was the case that the situation as depicted by Hol and Weijnen reflected the communis opinio until the publication of the Goeman-Taeldeman-Van Reenen data. One of science's goals is to test for 'truth' the generally accepted 'stories'.

Consider the option to use black-and-white symbols; these may help the map become clearer.

Just another example

Let's have another example: choose 'ft, cht' (or 'singular'), choose 'knecht' and fill out the query form like this: request: k omit: t. All dialect forms of 'knecht' start with k and a positive request is needed when a negative one is to be formulated: show all forms without t. It is quite a popular theme within (international) dialectology, the t being left out at the end of a word (t-deletion).
(Cf. also Weijnen's "Nederlandse Dialectkunde": map "CHT > CH".)

Now it is nice to compare this map with another: choose 'sentences', choose 'is de baas al thuis?', and search for: st. (We look for those places with pronunciation: 'is te baas' in stead of 'is de baas')
Adaptation of a sound to the features the sound immediately following ('to run ahead of things') is the usual case; now it appears that the localities which do it the reverse way (adaptation to what has been) overlap considerably with the localities of t-deletion, as a kind of border region between the two. What could this hint at?

To visualize overlap use the available → synthese-form.


Some notes for advanced users

When used to the interface of the query form now for some more complicated queries.

misc. | quick map vs first map | extra-linguistic | transcriber diff. | colorblinded | rework


* First of all, mind the choice foor 'and' or 'or' when applying more than one query string within one symbol. Either all four strings apply at once ('and') or at least one of them should do so ('or'); the latter being useful to vary contexts. Default choice is 'or'.

* Consider the use of black & white symbols.

* In order to specify word ending add a space, to specify the beginning of a word prefix a space.

* It is recommendable to first look through the list of transcriptions of a selected item. In doing so we learn more on the specifics of the current variation which helps to decide on what to expressly exclude from the query.
Take for example the word "huis". We should like to reconstruct the famous map by Kloeke and look for pronunciations 'hoes' and 'huus'. However, the diphtongue 'ui' often has 'uu' as a final sound. Final sounds of diphtongues are lightly spoken and thus end up between brackets. We add to our query for 'uu' the exclusion ("omit") of '(uu)'.

* Of course not all 'noise' can be excluded this way. Or unjustified exclusions prevented. To ensure an adequate map use the 'first map'-button. Beneath the resulting map a list of all data is presented per symbol with an option for correction of analysis. When corrected click 'second map' and your map will comply completely with your analysis. It certainly does not take much time to look through all data.

* Another problem is finding a character which may occur more than once within one transcription. In that case close in on the desired character by including some context in your query.
Take for example another map by Kloeke, the one on the spread of 'du' as a second person pronoun. (Note regarding the old map: 'du' is attested within the uncolored areas.) Now choose the category of "sentences". Choose the very first sentence "Je/ge moet die tak afzagen". Now it is immediately clear that a d- will often also occur with 'die'. To find the first d- (of 'du') we make use of the fact that initial sounds usually show little variation. Then it is clear also that almost every transcript will contain the m- of 'moet'. Our query then becomes: begin: d end: m. In other words: find a d before m.

* When selecting items on the basis of the ending of the stems you should realize it concerns the stem and not any realisation. For example, realisations like 'ngk' and 'nk' are not only to be found via selecting category "ng (+ k)"; a diminutive like 'schoentje' also has realisations with 'ngk' and 'nk' ('schoengke', schoenke').

'quick map' or 'first map'?

It will almost always be profitable and efficient to first create a 'quick map'. The results will indicate whether the query was rightly formulated or is in need of some adaptation. For ensuring a pure map choose next the option 'first map'. Beneath the resulting map you will find opportunity for correcting the assignment of symbols for the 'difficult' cases. Mind that the assignment of symbols is represented by whether a transcript complies with query a, b and/or c. (Default is 0, zero.) Leave all windows open in order to be able to return to your query form or list or map.

Mind that a 'first map' may take a minute (or two?). The building of the list for corrections takes some time. And with any option: the more places to be shown on the map the (slightly) longer it takes.

extra-linguistic information

It is possible to choose another base map for your dialect map through the option button in the left upper corner. It might be revealing, for example, to test whether correlations might exist between your data and for instance religious affiliation.

(The suggested correlation does not exist in our case, by the way. Yet examples are known of apparent correlations between the circumventing of taboo words like 'lightning' of 'thunder' and historical protestant affiliation of the appropriate areas.
(More in a coming publication by dr A. Goeman.)

(Tip: compare both catholic and protestant maps; it might be that localities find themselves right in between these areas.)

Of course it is the case that climate and soil determine the growth and characteristics of communities and thus may sometimes correlate with dialect differences. Of another type however are maps of communication routes like the one showing the maximum size of ships per river and channel. This map may correlate with the spread of certain features.

By choosing "none" all maps are emptied. When you choose another base map again this map will then be shown without the provinces underneath. The provinces will hint at where both maps do not match exactly, but a print with (or a look at) the thematic map might be clearer without the provinces appearing underneath.

A special feature is the option of putting in your own base map by entering its url in the appropriate input area below the map just above the lists. In order to prepare your image copy our basic base map from here. Take care to morph your map in such a way that
Westkapelle (I058p; Zeeland, west most Walcheren) has coordinates of circa 146px from the left and 359px from the top,
Vaals (Q222p; southern Limburg) 417 from the left, 495 from the top and
the top point of the province of Groningen 484 from the left, 16 from the top.
You should probably have to re-proportion your image (i.e. after an initial re-scaling) with a slightly greater resize of the height with respect to the width (e.g. 105 percent wider while 112 percent higher or 101 percent wider and 108 percent higher).

transcriber differences?

Click here for a map to check your results with when you are in doubt of possible (rare) transcriber differences surfacing in your data.

colour blind?

The colored symbols are meant to contrast sufficiently. The logic of the mixed colors shoukd be immediately obvious (red+yellow=orange, blue+yellow=green, blue+red=purple, all=black); adaptation for color blindedness meant giving up on this logic. When you should wish to present maps interpretable by colorblinded colleagues you may either use black & white symbols or upload a screen-image of your map at to have it daltonized (for maximum contrast when colorblind).


Next to the usual adaptations possible by means of adapting the HTML code or the pictures within the gif-directory (see saving maps) there is of course the option of reworking your map within a graphics application. A suggestion might be to save the map by means of a screen picture, to save a blank map as well, to open both in a graphics app, to make the blank map completely transparent and, finally, to draw an old-fashioned isogloss map within the blank map on the basis of the symbol map visible underneath. (This practice is found, for instance, in the "Klankgeografie" of the Brabantic Dictionaries.→ publications)

* Some more might follow. *



It is possible to add up two or more maps into one new map. This may come in handy for at least two reasons.

Firstly, when testing a feature against several items it may be insightful to add up all results in order to find a central area for this specific variation. For a certain feature may turn up in several words but in a central area it will turn up in most if not all words. The symbol of the new map will indicate the class of occurences it belongs to.

Secondly, when different themes are tested it may be insightful to see whether these themes cluster in which areas.

Beneath each map a field appears containing all results for all localities within the project: "a00,000,a00,a00". This line should be selected, copied and pasted in the 'synthesis'-form (or kept in a text document for later use). You will find the link to the synthesis form at the transcriptions' main page.

It is important that you copy the complete line (click in the text area and "select all"), otherwise the map will not be made.

Repeat these steps and fill out the synthesis form.

Important: indicate on the form for each line which of the results needs be used in the addition: a, b, c, ab, ac, bc of abc. As just mentioned, a, b, c refer to the query form. Thus, when you look into a recurring feature over several items it might be efficient to formulate each query in the same order.

Next decide which of both options you want.

Multiple maps around one theme (each map concerning a different item) may be added up into a map showing for each locality the number of times it showed in the separate maps. The assigned symbol represents one of five classes of frequencies. You may wish to manipulate the classes to highlight the lower or higher frequencies.
For this option check the "intensity" button.

Multiple maps on separate themes may be combined using symbol manipulation. Each row is represented by one of five features of a symbol, eg. bold vs light, vertical vs horizontal (or rounded). Five features (and their opposites) are considered to be the maximum without compromising the map's clarity.
Mind filling out thetext for the legends.
For this option check the "mix"-button.

Finally, click the "show map"-button to call for the new map.

Each symbol consists of five characteristics: form, color, filling, markedness, size. Naturally a symbol has form and color. The other three features may or may not be present. Combining features means that when a locality turns up with, say 'red' but is not included in the selection assigned with 'filling' it receives the default value 'not filled'. Should the locality turn up in both selections it's symbol would have 'red' and a filling.

The legend will show each of the resulting symbols but will also show the used features separately.

A trick might be to set the symbol feature for a row which is not used in order to use its default feature. For example, when the row belonging to the 'filled/not filled'-feature is set to 'not filled', the default value for all other symbols will be 'filled'. This way you may have more strongly visible symbols on your map.

Click here for the social variables pertaining to the speakers in the format for use in the 'synthesis' form.

Click here for more on the alternative to these online options, the off-line Kloeketabel application.


Saving maps

In order to store a map please save as
in Explorer: "HTML-source" (not "web archive")
in Netscape / Mozilla: "webpage complete".
To view the map again (offline) using Explorer you need the folder / directory with all symbols and ground maps. This folder is to be downloaded here (zip archive). Keep folder and saved HTML-file at the same level (i.e. within the same directory; not one inside the other) and open the file in your browser.

Of course you can adapt the HTML-source code after your own liking or design your own symbols or ground maps in a paint program. Mind to keep the file name of the symbol image the same as is used within the source file.

Your maps (with reference) may also be sent to us to be published at the website of the Morphology Atlas (MAND). A variety of maps may thus become available.

(As yet it is not possible to save maps using Safari 1 / OS X 2. Take a snapshot. Safari 2 / OS X 3 has not been tested yet.)




Click here for an assignment as is given each year to students at the Vrije Universiteit (Free University at Amsterdam) by Eric Akkerman, designed with his cooperation by Boudewijn van den Berg.


Transcriptions per place

In order to view the transcriptions of all items of one and the same place, follow the link to 'transcriptions per place' and make your choice.

Data presentation knows three formats:

  • ordered by stem endings (final consonant),
  • ordered by word class (plural, diminutive, etc.)
  • alphabetical per word type (in order of elicitation; default format).
(The first two options mirror the tables of the selection screen 'transcriptions per item'.)

Coming with the first two formats is the option to include each main vowel's west-germanic origin the ordering has been based on. The first presentation will thus represent an almost allround sound grammar of the particular dialect, the second presentation (ordered by word class) represents a fair part of the particular dialect's word grammar.

The third option lends itself for a quick scan or, for that matter, for copying all data into a database of your own.

Copying all of a presentation proceeds via 'select all', 'copy' and a 'paste' into your text file.

A suggestion might be to design a work booklet with introduction (some examples of which are in preparation for this site), notes on the notation of pronunciation, characteristics of the speaker, the list of items & pronunciation (edited as you deem fit), room for notes, some maps, of a characteristically pronounced item or a synthesis-map of (say) five comparable items together.

(The CD-ROM allows for searches on several items in one go, thus skipping the need for copy and paste within the synthesis-form.)


Creative Commons License   GNU

Creative Commons | GNU-GPL

In the use in any form towards others of data or analyses based on these data, always mention: for Dutch data: "Meertens Instituut-KNAW (Goeman, Van Reenen)"; for Flemish data: "University of Ghent (Taeldeman)"; for Frisian data: "Fryske Akademy-KNAW".

In the use in any form towards others of the KIPA-encoding or of the simplified pronunciation spelling, always mention: "Goeman, Van Reenen, Van den Berg (Meertens Instituut-KNAW)".

Data, metadata, geographical codes and coordinates are not to be used for commercial purposes.

In the use in any form towards others of applications as present or mentioned at pages of MAND, GTRP and Cartography, always mention: "Van den Berg (Meertens Instituut-KNAW)".

All pages, scripts, looks and concepts (also those of the CD-ROM and applications, including renderings other than KIPA) are composed by Boudewijn van den Berg with the exceptions of a conversion script by Maarten van der Peet, the layout canvas of the Institute, the etymology text (by Ton Goeman), the slide-show presentations (in cooperation) and of course the scanned articles by different authors.


Creative Commons