Methodology

Corpus

The corpus for this research came from a public art project called Humans of Amman. Humans of Amman captures photographs of people around Amman, Jordan and adds captions given by the subjects of the photographs. These captions are what were used in the research to analyze the language these Jordanians use.

Thanks to all the people behind the Humans of Amman project!

To learn more about the project, you can email them at info@humansofamman.net, visit their website here, or check out their Facebook page.

The Grammar Variations

The research project focuses on three main grammar variations; gender agreement, verb prefixes, and elision.

I looked at gender agreement among nouns and adjectives. In Arabic, some nouns are feminine and others masculine, with most feminine nouns ending in a specific letter marker, called a tarma-butta. Nouns with this feminine marker or that are non-human and plural must have their matching adjectives also contain the feminine letter marker. I marked nouns and adjectives that agreed, whether masculine or feminine, with a ‘y’ attribute and those that didn’t agree with a ‘n’. This gender agreement also exists very similarly in Modern Standard Arabic but I choose to test it to see that if in dialect speakers would get lazy and at times forget to agree in gender.

Elision refers to the omission of a sound of syllable when speaking. In dialect, there’s a tendency to smash together some prepositions with their preceding nouns. The point in marking these prepositions was too see when elision did or did not take place and how the presence of absence of a definite marker on the following noun affected this elision.

There is a letter ‘b’ prefix that is often put in front of verbs in dialectal Arabic. Its usage varies and there’s no concrete rule about when or on what verbs it should and shouldn’t be used. It can often be used to show emphasis on a particular verb, especially when there is more than one verb in a row. In addition, it’s in most cases not used when it follows another verb, whether it’s a plain verb, auxiliary verb, or a participle.

Hypotheses

Gender

There will be few gender disagreements in the excerpts because gender is so ingrained in the language of Arabic and is not often ignored even in dialect. Therefore, I hypothesize that speakers will respect gender agreement the large majority of the time even when speaking in dialect.

Prepositions

When there is the preposition 'a'la' followed by a noun, whether with or without the definite marker, there will be greater instances of elision in which the two words are 'connected'.

Verbs

Overall, there will be more verbs without a ‘b’ prefix than with. However, I hypothesize these differences to be closer in plain verbs as opposed to the other marked verbs, ones that are marked as having come after a different tense. In the other verbs, I hypothesize that there will be marginally few verbs that contain a ‘b’ prefix.

Explaining Schema and Mark-Up

Element: Gender

Attribute: Agree

Attribute Value: "y" | "n"

Element: Verb

Attribute: Prefix

Attribute Value: "y" | "n" | "y 3m" | "n 3m" | "n future" | "y future" | "n neg" | "y neg" | "n past" | "y past" | "y 2nd" | "n 2nd" | "n knt" | "y knt" | "n part" | "y part" | "n aux" | "y aux"

3m verb = a verb marker that turns the verb after it into a gerund, or -ing verb

knt verb = 'was', a verb that marks the to be verb in the past tense

Element: Preposition

Attribute: Connect

Attribute Value: "y def" | "n def" | "y indef" | "n indef"

Things to Consider

Because we used a public project as part of our corpus, there are a few things we should consider that could effect the outcome of the analysis.

  1. These excerpts are transcripts of someone speaking so while the transcriber did their best to stick to the original spoken Arabic, there may be mistakes in transcription that could potentially throw off the outcome of the research.
  2. These excerpts were taken from people around the city of Amman. Therefore, they are not all necessarily native speakers of Jordanian Arabic. They could possibly be from other areas of the Arabic speaking world or non-native speakers and thus their speaking may not be directly in line with mainstream Jordanian dialect.
  3. I, the researcher, am not a native Arabic speaker. I've studied four years worth of Arabic at the University and have spent time living in Amman, Jordan and though I did my best to choose a research question and research goals that fit my level of Arabic knowledge, there may be circumstances where I misunderstood something that would lead to skewing the data a bit.

A Note About Formatting

From attempting to undergo a research project in a non-Western language, I realized just how biased towards left-to-right formatted languages computers could be. The mark-up of this project underwent a lot of delays in the beginning (and a lot of frustration through out) as I attempted to figure out how to mark up the Arabic text with English elements and attributes. In order to do this in Oxygen, the class’s preferred XML Editor, I had to carefully mark the Arabic text up by elements and then switch to the Author mode in order to add the attributes. If I made a mistake and needed to change or erase an attribute or an element, I had to be really careful about altering the text because of the way the left-to-right nature of the English text and the right-to-left of the Arabic text worked together.

Overall, I made it work and the formatting issues didn’t disrupt the research in any way other than general frustration. However, it’s an interesting observation and consideration to take in if you are attempting to work with different language formats in the future. If you like to take a look at my mark up and especially how the Arabic mixed with the English in the XML document and also on the Analysis pages, you can do as at my Github here.

Creative Commons License
Jordanian Arabic by Amber Montgomery is licensed under a Creative Commons Attribution 4.0 International License.