Looking at the ezāfe construction from a linguistic perspective, part 2: a fresh approach to ezāfe analysis.

In the last blog post (part 1) we looked at some different examples of اضافه ezāfe in Persian, and learned how the different realisations of the ezāfe construction are categorised within Persian grammar.

In part 2, I want to do something a little risqué (linguistically-speaking) and pick apart the traditional way of analysing ezāfe. I feel like this is a good time to write a little disclaimer: I am by no means an expert and my opinion should be taken with a grain of salt.

My issue with traditional analyses of ezāfe is that they rely on predominantly semantic distinctions, rather than on grammatical ones. There must be an overarching grammatical explanation for all of the different ezāfe constructions we come across in Persian. Basically, what allows that pesky little linking -e to pop up all over the place in Persian phrases?

In this blog, I will propose that we can classify the ezāfe construction as being a kind of head-marking agreement suffix. What does this mean and how did I come to this conclusion? Well I’m glad you asked…

Types of ezāfe construction (انواع ترکیب اضافی)

Grammatically-speaking, there are a lot of different ways in which we can use ezāfe in Persian. Here are the 5 main ones I found:

1. Noun + noun (اسم به اضافۀ اسم)

E.g. خانۀ نرگس khāne-ye Narges ‘Narguess’ house’
or باغِ من bāgh-e man ‘my garden’

(All of the examples we looked at in part 1 fall into this category. In traditional Persian grammar, the term ‘ezāfe’ is only used to describe the linking of two nouns together.)

2. Noun + adjective (اسم به اضافۀ صفت)

E.g. خانۀ بزرگ khāne-ye bozorg ‘the big house’

(This is not classified as ezāfe by Persian grammarians. However, I would argue that in terms of grammar, it can be described as being exactly the same syntactic phenomenon.)

It is also very common to come across ezāfe ‘chains’, where nouns are linked to adjectives which are linked to more nouns with more adjectives – there is technically no limit to how long we can make one of these ‘chains’. For example:

باغِ گُلِ زیبای دوستِ صَمیمیِ من

Bāgh-e gol-e zibā-ye dust-e samimi-ye man

‘My dear friend’s beautiful flower garden’

(Noun + noun + adjective + noun + adjective + pronoun!)

Here’s another even longer example:

کتابِ شاگردِ کلاسِ سومِ دبیرستانِ محلِ دوستِ قدیمیِ پدرِ من

Ketāb-e shāgerd-e kelās-e sevom-e dabirestān-e mahall-e dust-e qadimi-ye pedar-e man

‘My father’s old friend’s neighbourhood highschool’s third grade student’s book’

(Noun + noun + noun + number + noun + noun + noun + adjective + noun + pronoun!)

3. Noun + reduced relative clause (اسم به اضافۀ بندِ فرعی)

E.g. معلمانِ از تعطیلات برگشته mu’allemān-e az ta’tilāt bar-gashte ‘the teachers [who] returned from holiday’

(This type of ezāfe can be replaced by using the word که ke instead. E.g. معلمانی که از تعطیلات برگشته mu’allemān-i ke az ta’tilāt bar-gashte.)

4. Verbal infinitive or nominalised verb phrase + noun (اسمِ مصدر به اضافۀ اسم)

E.g. پختنِ کیک pokhtan-e keyk ‘baking the cake’

5. Preposition + noun (حرفِ افزونه به اضافۀ اسم)

E.g. بیرونِ خانۀ birun-e khāne ‘outside the house’

(This is a tricky one to categorise, as بیرون birun here is technically a noun: ‘the outside’. However, I have included it for the sake of comparison with English grammar.)

Features of the ezāfe

By looking at the examples above we can begin to notice some patterns emerging. The مُضاف mozāf ‘head’ is always the first word in a Persian ezāfe chain, and the ‘dependants’ always come afterwards. Also, it is important to note that the ezāfe chain will never begin before the head word. This is the reason why modifiers such as این in ‘this’, آن ān ‘that’, هر har ‘each/every’, هیچ hich ‘no/any’, یک yek ‘one’, and دو do ‘two’ never require ezāfe, because they are always placed before the head.

An interesting case which illustrates this rule is the word تنها tanhā, which changes meaning depending on whether it is placed before or after the head. Before the head as a modifier (therefore without ezāfe), it means ‘only’, such as تنها مردی که tanhā mard-i ke ‘the only man that…’. After the head as an adjective (therefore requiring ezāfe), it means ‘lonely’ or ‘alone’, such as مردِ تنها mard-e tanhā ‘the lonely man’.

Development

Ezāfe is considered to be unique to Persian and to the Iranian language family. So how did this construction develop?

It is believed that ezāfe originated from a relative pronoun, namely the word ī in Pahlavi (Middle Persian), meaning ‘which’. So in Middle Persian, a phrase such as mard ī tanīhā would have meant something like ‘the man which [is] alone/lonely’. Over time, this construction was reanalysed to become the modern ezāfe construction (the head-dependant relationship) which we have already described above: مردِ تنها mard-e tanhā ‘the lonely man’.

So if we go back to No. 3 on our list of ways in which ezāfe can be used, ‘noun + reduced relative clause’, we can say that this was actually the earliest and original form of the ezāfe, stemming from the relative pronoun ‘which’.

Linguistic analysis

Finally, how can we analyse ezāfe constructions from a linguistic perspective? Many people will argue that ezāfe is a grammatical case, however this would be an oversimplification of what is actually going on here. It is easy to call ezāfe a grammatical case and then close the door on the subject, never to be opened again unless you are an argumentative linguist. However, just because the Arabic إضافة ʾiḍāfa construction is clearly an expression of the genitive case, that does not mean the same is true for Persian. The Persian ezāfe can’t be an expression of grammatical case because we can also use ezāfe for reduced relative clauses, a grammatical concept which surpasses the notion of case altogether. There must be another linguistic explanation for what is causing all of these varied ezāfe relations.

We are left with two options: either it’s grammatical agreement or it’s a head-marking suffix.

The argument for ezāfe being some kind of agreement is backed up by evidence from the Kurdish languages, which display more widespread agreement than Modern Persian does. Let’s take a look at an example from Kurmanji:

مالا مَزینا نَرمینه

Māl-ā mazin-ā Narmina

‘Narmin’s big house’

The above Kurmanji phrase has gender agreement: the suffix -ā is used to express ezāfe with feminine agreement. Therefore, one can say that even in a language without grammatical gender, such as Modern Persian, ezāfe can still be analysed as being a form of agreement.

The other argument is that ezāfe is a simple head-marking suffix. This is the most straight-forward and painless assessment we can make, especially as we have already clearly seen how ezāfe links together a head with its dependant(s) in a ‘chain’, in order to mark the head of the chain. The role of the ezāfe is simply to delineate which specific words are the dependants for a given head.

Personally, I believe that both of these arguments can be correct. This is how I arrived at my long-winded linguistic title: ezāfe is a head-marking agreement suffix.

Clearly this is something which will continue to be debated for years to come. Whatever you, or the bickering linguists think about all of this, at least you can now rest assured that you know more about the ezāfe construction than you ever really wanted to know in the first place! If you disagree with my analysis, or have your own one to propose, I would love to hear all about it in the comments!

– Sam